Home/ai-coding-skills
2
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

feat: add pi reviewer support to workflow variants

Browse Source
This commit is contained in:
Stefano Fiorini
2026-04-23 21:03:45 -05:00
parent ce4746b769
commit 231a66f2b1
26 changed files with 512 additions and 20 deletions
Download Patch File Download Diff File Expand all files Collapse all files
skills/create-plan/claude-code/SKILL.md
+38 -1
View File
@@ -33,6 +33,19 @@ If any dependency is missing, stop immediately and return:
### Phase 3: Configure Reviewer
Reviewer CLI: `codex`, `claude`, `cursor`, `opencode`, `pi`, or `skip`.
If `REVIEWER_CLI=pi`, verify the Pi reviewer binary before entering the review loop:
```bash
pi --version
```
For shorthand `pi/<pi-model-name>`, split only on the first slash when the prefix is exactly `pi`; store the complete remainder in `REVIEWER_MODEL`. Examples: `pi/claude-opus-4-7` -> `claude-opus-4-7`, `pi/anthropic/claude-opus-4-7` -> `anthropic/claude-opus-4-7`, and `pi/openrouter/anthropic/claude-opus-4-7` -> `openrouter/anthropic/claude-opus-4-7`.
When `REVIEWER_CLI=pi`, the reviewer model is configured independently from the model running this workflow. If the model/provider is unavailable, surface helper stderr/status and use `pi --list-models [search]` to inspect configured models.
If the user has already specified a reviewer CLI and model (e.g., "create a plan, review with codex o4-mini"), use those values. Otherwise, ask:
1. **Which CLI should review the plan?**
@@ -152,6 +165,18 @@ Write the reviewer invocation to `/tmp/plan-review-${REVIEW_ID}.sh` as a bash sc
set -euo pipefail
```
**If `REVIEWER_CLI` is `pi`:**
Fresh call every round (Pi reviewer calls do not use session resume):
```bash
pi --no-session --no-skills --no-prompt-templates --no-extensions --no-context-files \
--model "$REVIEWER_MODEL" \
--tools read,grep,find,ls \
-p "Read the file /tmp/plan-${REVIEW_ID}.md and review. Return exactly the required ## Summary, ## Findings, and ## Verdict structure."
```
**If `REVIEWER_CLI` is `codex`:**
```bash
@@ -281,7 +306,7 @@ jq -r '.result' /tmp/plan-review-${REVIEW_ID}.json > /tmp/plan-review-${REVIEW_I
```
- If `REVIEWER_CLI=codex`, extract `CODEX_SESSION_ID` from `/tmp/plan-review-${REVIEW_ID}.runner.out` after the helper or fallback run. If the review text is only in `.runner.out`, move or copy the actual review body into `/tmp/plan-review-${REVIEW_ID}.md` before verdict parsing.
- If `REVIEWER_CLI=claude`, promote stdout captured by the helper or fallback runner into the markdown review file:
- If `REVIEWER_CLI=claude` or `REVIEWER_CLI=pi`, promote stdout captured by the helper or fallback runner into the markdown review file:
```bash
cp /tmp/plan-review-${REVIEW_ID}.runner.out /tmp/plan-review-${REVIEW_ID}.md
@@ -331,6 +356,18 @@ If a revision contradicts the user's explicit requirements, skip it and note it
Rewrite `/tmp/plan-review-${REVIEW_ID}.sh` for the next round. The script should contain the reviewer invocation only; do not run it directly.
**If `REVIEWER_CLI` is `pi`:**
Fresh call with prior-round context (Pi reviewer calls do not use session resume):
```bash
pi --no-session --no-skills --no-prompt-templates --no-extensions --no-context-files \
--model "$REVIEWER_MODEL" \
--tools read,grep,find,ls \
-p "You previously reviewed this plan and requested revisions. Read the updated payload at /tmp/plan-${REVIEW_ID}.md and re-review using the same ## Summary, ## Findings, and ## Verdict structure."
```
**If `REVIEWER_CLI` is `codex`:**
Resume the existing session:
skills/create-plan/codex/SKILL.md
+38 -1
View File
@@ -58,6 +58,19 @@ If any dependency is missing, stop and return:
### Phase 3: Configure Reviewer
Reviewer CLI: `codex`, `claude`, `cursor`, `opencode`, `pi`, or `skip`.
If `REVIEWER_CLI=pi`, verify the Pi reviewer binary before entering the review loop:
```bash
pi --version
```
For shorthand `pi/<pi-model-name>`, split only on the first slash when the prefix is exactly `pi`; store the complete remainder in `REVIEWER_MODEL`. Examples: `pi/claude-opus-4-7` -> `claude-opus-4-7`, `pi/anthropic/claude-opus-4-7` -> `anthropic/claude-opus-4-7`, and `pi/openrouter/anthropic/claude-opus-4-7` -> `openrouter/anthropic/claude-opus-4-7`.
When `REVIEWER_CLI=pi`, the reviewer model is configured independently from the model running this workflow. If the model/provider is unavailable, surface helper stderr/status and use `pi --list-models [search]` to inspect configured models.
If the user has already specified a reviewer CLI and model (e.g., "create a plan, review with claude sonnet"), use those values. Otherwise, ask:
1. **Which CLI should review the plan?**
@@ -175,6 +188,18 @@ Write the reviewer invocation to `/tmp/plan-review-${REVIEW_ID}.sh` as a bash sc
set -euo pipefail
```
**If `REVIEWER_CLI` is `pi`:**
Fresh call every round (Pi reviewer calls do not use session resume):
```bash
pi --no-session --no-skills --no-prompt-templates --no-extensions --no-context-files \
--model "$REVIEWER_MODEL" \
--tools read,grep,find,ls \
-p "Read the file /tmp/plan-${REVIEW_ID}.md and review. Return exactly the required ## Summary, ## Findings, and ## Verdict structure."
```
**If `REVIEWER_CLI` is `codex`:**
```bash
@@ -304,7 +329,7 @@ jq -r '.result' /tmp/plan-review-${REVIEW_ID}.json > /tmp/plan-review-${REVIEW_I
```
- If `REVIEWER_CLI=codex`, extract `CODEX_SESSION_ID` from `/tmp/plan-review-${REVIEW_ID}.runner.out` after the helper or fallback run. If the review text is only in `.runner.out`, move or copy the actual review body into `/tmp/plan-review-${REVIEW_ID}.md` before verdict parsing.
- If `REVIEWER_CLI=claude`, promote stdout captured by the helper or fallback runner into the markdown review file:
- If `REVIEWER_CLI=claude` or `REVIEWER_CLI=pi`, promote stdout captured by the helper or fallback runner into the markdown review file:
```bash
cp /tmp/plan-review-${REVIEW_ID}.runner.out /tmp/plan-review-${REVIEW_ID}.md
@@ -354,6 +379,18 @@ If a revision contradicts the user's explicit requirements, skip it and note it
Rewrite `/tmp/plan-review-${REVIEW_ID}.sh` for the next round. The script should contain the reviewer invocation only; do not run it directly.
**If `REVIEWER_CLI` is `pi`:**
Fresh call with prior-round context (Pi reviewer calls do not use session resume):
```bash
pi --no-session --no-skills --no-prompt-templates --no-extensions --no-context-files \
--model "$REVIEWER_MODEL" \
--tools read,grep,find,ls \
-p "You previously reviewed this plan and requested revisions. Read the updated payload at /tmp/plan-${REVIEW_ID}.md and re-review using the same ## Summary, ## Findings, and ## Verdict structure."
```
**If `REVIEWER_CLI` is `codex`:**
Resume the existing session:
skills/create-plan/cursor/SKILL.md
+38 -1
View File
@@ -59,6 +59,19 @@ If any dependency is missing, stop and return:
### Phase 3: Configure Reviewer
Reviewer CLI: `codex`, `claude`, `cursor`, `opencode`, `pi`, or `skip`.
If `REVIEWER_CLI=pi`, verify the Pi reviewer binary before entering the review loop:
```bash
pi --version
```
For shorthand `pi/<pi-model-name>`, split only on the first slash when the prefix is exactly `pi`; store the complete remainder in `REVIEWER_MODEL`. Examples: `pi/claude-opus-4-7` -> `claude-opus-4-7`, `pi/anthropic/claude-opus-4-7` -> `anthropic/claude-opus-4-7`, and `pi/openrouter/anthropic/claude-opus-4-7` -> `openrouter/anthropic/claude-opus-4-7`.
When `REVIEWER_CLI=pi`, the reviewer model is configured independently from the model running this workflow. If the model/provider is unavailable, surface helper stderr/status and use `pi --list-models [search]` to inspect configured models.
If the user has already specified a reviewer CLI and model (e.g., "create a plan, review with codex o4-mini"), use those values. Otherwise, ask:
1. **Which CLI should review the plan?**
@@ -181,6 +194,18 @@ Write the reviewer invocation to `/tmp/plan-review-${REVIEW_ID}.sh` as a bash sc
set -euo pipefail
```
**If `REVIEWER_CLI` is `pi`:**
Fresh call every round (Pi reviewer calls do not use session resume):
```bash
pi --no-session --no-skills --no-prompt-templates --no-extensions --no-context-files \
--model "$REVIEWER_MODEL" \
--tools read,grep,find,ls \
-p "Read the file /tmp/plan-${REVIEW_ID}.md and review. Return exactly the required ## Summary, ## Findings, and ## Verdict structure."
```
**If `REVIEWER_CLI` is `codex`:**
```bash
@@ -316,7 +341,7 @@ jq -r '.result' /tmp/plan-review-${REVIEW_ID}.json > /tmp/plan-review-${REVIEW_I
```
- If `REVIEWER_CLI=codex`, extract `CODEX_SESSION_ID` from `/tmp/plan-review-${REVIEW_ID}.runner.out` after the helper or fallback run. If the review text is only in `.runner.out`, move or copy the actual review body into `/tmp/plan-review-${REVIEW_ID}.md` before verdict parsing.
- If `REVIEWER_CLI=claude`, promote stdout captured by the helper or fallback runner into the markdown review file:
- If `REVIEWER_CLI=claude` or `REVIEWER_CLI=pi`, promote stdout captured by the helper or fallback runner into the markdown review file:
```bash
cp /tmp/plan-review-${REVIEW_ID}.runner.out /tmp/plan-review-${REVIEW_ID}.md
@@ -364,6 +389,18 @@ If a revision contradicts the user's explicit requirements, skip it and note it
Rewrite `/tmp/plan-review-${REVIEW_ID}.sh` for the next round. The script should contain the reviewer invocation only; do not run it directly.
**If `REVIEWER_CLI` is `pi`:**
Fresh call with prior-round context (Pi reviewer calls do not use session resume):
```bash
pi --no-session --no-skills --no-prompt-templates --no-extensions --no-context-files \
--model "$REVIEWER_MODEL" \
--tools read,grep,find,ls \
-p "You previously reviewed this plan and requested revisions. Read the updated payload at /tmp/plan-${REVIEW_ID}.md and re-review using the same ## Summary, ## Findings, and ## Verdict structure."
```
**If `REVIEWER_CLI` is `codex`:**
Resume the existing session:
skills/create-plan/opencode/SKILL.md
+38 -1
View File
@@ -64,6 +64,19 @@ If the user has already specified a reviewer CLI and model (e.g., "create a plan
Store the chosen `REVIEWER_CLI`, `REVIEWER_MODEL`, and `MAX_ROUNDS` for Phase 7 (Iterative Plan Review).
Reviewer CLI: `codex`, `claude`, `cursor`, `opencode`, `pi`, or `skip`.
If `REVIEWER_CLI=pi`, verify the Pi reviewer binary before entering the review loop:
```bash
pi --version
```
For shorthand `pi/<pi-model-name>`, split only on the first slash when the prefix is exactly `pi`; store the complete remainder in `REVIEWER_MODEL`. Examples: `pi/claude-opus-4-7` -> `claude-opus-4-7`, `pi/anthropic/claude-opus-4-7` -> `anthropic/claude-opus-4-7`, and `pi/openrouter/anthropic/claude-opus-4-7` -> `openrouter/anthropic/claude-opus-4-7`.
When `REVIEWER_CLI=pi`, the reviewer model is configured independently from the model running this workflow. If the model/provider is unavailable, surface helper stderr/status and use `pi --list-models [search]` to inspect configured models.
### Phase 5: Design (REQUIRED SUB-SKILL)
Use OpenCode's native skill tool to load:
@@ -169,6 +182,18 @@ Write the reviewer invocation to `/tmp/plan-review-${REVIEW_ID}.sh` as a bash sc
set -euo pipefail
```
**If `REVIEWER_CLI` is `pi`:**
Fresh call every round (Pi reviewer calls do not use session resume):
```bash
pi --no-session --no-skills --no-prompt-templates --no-extensions --no-context-files \
--model "$REVIEWER_MODEL" \
--tools read,grep,find,ls \
-p "Read the file /tmp/plan-${REVIEW_ID}.md and review. Return exactly the required ## Summary, ## Findings, and ## Verdict structure."
```
**If `REVIEWER_CLI` is `codex`:**
```bash
@@ -298,7 +323,7 @@ jq -r '.result' /tmp/plan-review-${REVIEW_ID}.json > /tmp/plan-review-${REVIEW_I
```
- If `REVIEWER_CLI=codex`, extract `CODEX_SESSION_ID` from `/tmp/plan-review-${REVIEW_ID}.runner.out` after the helper or fallback run. If the review text is only in `.runner.out`, move or copy the actual review body into `/tmp/plan-review-${REVIEW_ID}.md` before verdict parsing.
- If `REVIEWER_CLI=claude`, promote stdout captured by the helper or fallback runner into the markdown review file:
- If `REVIEWER_CLI=claude` or `REVIEWER_CLI=pi`, promote stdout captured by the helper or fallback runner into the markdown review file:
```bash
cp /tmp/plan-review-${REVIEW_ID}.runner.out /tmp/plan-review-${REVIEW_ID}.md
@@ -346,6 +371,18 @@ If a revision contradicts the user's explicit requirements, skip it and note it
Rewrite `/tmp/plan-review-${REVIEW_ID}.sh` for the next round. The script should contain the reviewer invocation only; do not run it directly.
**If `REVIEWER_CLI` is `pi`:**
Fresh call with prior-round context (Pi reviewer calls do not use session resume):
```bash
pi --no-session --no-skills --no-prompt-templates --no-extensions --no-context-files \
--model "$REVIEWER_MODEL" \
--tools read,grep,find,ls \
-p "You previously reviewed this plan and requested revisions. Read the updated payload at /tmp/plan-${REVIEW_ID}.md and re-review using the same ## Summary, ## Findings, and ## Verdict structure."
```
**If `REVIEWER_CLI` is `codex`:**
Resume the existing session:
skills/create-plan/pi/SKILL.md
+8
View File
@@ -85,6 +85,14 @@ Reviewer CLI: `codex`, `claude`, `cursor`, `opencode`, `pi`, or `skip`
Store `REVIEWER_CLI`, `REVIEWER_MODEL`, and `MAX_ROUNDS` for the review loop.
If `REVIEWER_CLI=pi`, verify the Pi reviewer binary before entering the review loop:
```bash
pi --version
```
For shorthand `pi/<pi-model-name>`, split only on the first slash when the prefix is exactly `pi`; store the complete remainder in `REVIEWER_MODEL`. Examples: `pi/claude-opus-4-7` -> `claude-opus-4-7`, `pi/anthropic/claude-opus-4-7` -> `anthropic/claude-opus-4-7`, and `pi/openrouter/anthropic/claude-opus-4-7` -> `openrouter/anthropic/claude-opus-4-7`.
When `REVIEWER_CLI=pi`, the reviewer model is configured independently from the pi model running this workflow. Use any configured pi model string, including provider-qualified model IDs. If the reviewer model or provider is unavailable, surface the review helper stderr/status and ask for a configured model; use `pi --list-models [search]` to inspect configured models.
The pi reviewer command rendered into `/tmp/plan-review-${REVIEW_ID}.sh` must be isolated and read-only:
skills/do-task/claude-code/SKILL.md
+27 -2
View File
@@ -111,6 +111,19 @@ If the user has already specified a reviewer CLI and model (e.g., "do task X, re
Store `REVIEWER_CLI`, `REVIEWER_MODEL`, and `MAX_ROUNDS` for Phases 5 and 8.
Reviewer CLI: `codex`, `claude`, `cursor`, `opencode`, `pi`, or `skip`.
If `REVIEWER_CLI=pi`, verify the Pi reviewer binary before entering the review loop:
```bash
pi --version
```
For shorthand `pi/<pi-model-name>`, split only on the first slash when the prefix is exactly `pi`; store the complete remainder in `REVIEWER_MODEL`. Examples: `pi/claude-opus-4-7` -> `claude-opus-4-7`, `pi/anthropic/claude-opus-4-7` -> `anthropic/claude-opus-4-7`, and `pi/openrouter/anthropic/claude-opus-4-7` -> `openrouter/anthropic/claude-opus-4-7`.
When `REVIEWER_CLI=pi`, the reviewer model is configured independently from the model running this workflow. If the model/provider is unavailable, surface helper stderr/status and use `pi --list-models [search]` to inspect configured models.
### Phase 4: Initialize Plan Workspace
**PLAN MODE CHECK:** If currently in plan mode:
@@ -344,7 +357,7 @@ This subroutine is invoked twice per `do-task` run: once in Phase 5 (`REVIEW_KIN
| `REVIEW_ID` | 8-char hex (from `uuidgen`); reused across rounds of the same loop |
| `PAYLOAD_PATH` | `/tmp/do-task-${REVIEW_KIND}-${REVIEW_ID}.md` |
| `PROMPT_TEMPLATE` | `PLAN_REVIEW_PROMPT` or `IMPL_REVIEW_PROMPT` |
| `REVIEWER_CLI` | `codex` \| `claude` \| `cursor` \| `opencode` |
| `REVIEWER_CLI` | `codex` \| `claude` \| `cursor` \| `opencode` \| `pi` |
| `REVIEWER_MODEL` | Model name |
| `MAX_ROUNDS` | Default 10 |
| `SESSION_ID_VAR` | `CODEX_PLAN_SESSION_ID` \| `CODEX_IMPL_SESSION_ID` \| `CURSOR_PLAN_SESSION_ID` \| `CURSOR_IMPL_SESSION_ID` \| `OPENCODE_PLAN_SESSION_ID` \| `OPENCODE_IMPL_SESSION_ID` |
@@ -437,6 +450,18 @@ Write the reviewer invocation to `/tmp/do-task-${REVIEW_KIND}-review-${REVIEW_ID
set -euo pipefail
```
**If `REVIEWER_CLI` is `pi`:**
Fresh call every round (Pi reviewer calls do not use session resume):
```bash
pi --no-session --no-skills --no-prompt-templates --no-extensions --no-context-files \
--model "$REVIEWER_MODEL" \
--tools read,grep,find,ls \
-p "Read the file /tmp/do-task-${REVIEW_KIND}-${REVIEW_ID}.md and review. Return exactly the required ## Summary, ## Findings, and ## Verdict structure."
```
**If `REVIEWER_CLI` is `codex`:**
Round 1 — fresh `codex exec`:
@@ -633,7 +658,7 @@ After the command completes:
cp /tmp/do-task-${REVIEW_KIND}-review-${REVIEW_ID}.runner.out \
/tmp/do-task-${REVIEW_KIND}-review-${REVIEW_ID}.md
```
- `claude`: promote `.runner.out` into the `.md` file:
- `claude` or `pi`: promote `.runner.out` into the `.md` file:
```bash
cp /tmp/do-task-${REVIEW_KIND}-review-${REVIEW_ID}.runner.out \
/tmp/do-task-${REVIEW_KIND}-review-${REVIEW_ID}.md
skills/do-task/claude-code/templates/task-plan.md
+1 -1
View File
@@ -9,7 +9,7 @@
| Created | YYYY-MM-DD |
| Slug | YYYY-MM-DD-<slug> |
| Runtime | claude-code |
| Reviewer CLI | codex \| claude \| cursor \| opencode |
| Reviewer CLI | codex \| claude \| cursor \| opencode \| pi |
| Reviewer Model | <model> |
| MAX_ROUNDS | 10 |
| Branch Strategy | current-branch \| worktree |
skills/do-task/codex/SKILL.md
+27 -2
View File
@@ -133,6 +133,19 @@ If the user has already specified a reviewer CLI and model (e.g., "do task X, re
Store `REVIEWER_CLI`, `REVIEWER_MODEL`, and `MAX_ROUNDS` for Phases 5 and 8.
Reviewer CLI: `codex`, `claude`, `cursor`, `opencode`, `pi`, or `skip`.
If `REVIEWER_CLI=pi`, verify the Pi reviewer binary before entering the review loop:
```bash
pi --version
```
For shorthand `pi/<pi-model-name>`, split only on the first slash when the prefix is exactly `pi`; store the complete remainder in `REVIEWER_MODEL`. Examples: `pi/claude-opus-4-7` -> `claude-opus-4-7`, `pi/anthropic/claude-opus-4-7` -> `anthropic/claude-opus-4-7`, and `pi/openrouter/anthropic/claude-opus-4-7` -> `openrouter/anthropic/claude-opus-4-7`.
When `REVIEWER_CLI=pi`, the reviewer model is configured independently from the model running this workflow. If the model/provider is unavailable, surface helper stderr/status and use `pi --list-models [search]` to inspect configured models.
### Phase 4: Initialize Plan Workspace
Codex has no plan-mode concept; there is no plan-mode guard here.
@@ -363,7 +376,7 @@ This subroutine is invoked twice per `do-task` run: once in Phase 5 (`REVIEW_KIN
| `REVIEW_ID` | 8-char hex (from `uuidgen`); reused across rounds of the same loop |
| `PAYLOAD_PATH` | `/tmp/do-task-${REVIEW_KIND}-${REVIEW_ID}.md` |
| `PROMPT_TEMPLATE` | `PLAN_REVIEW_PROMPT` or `IMPL_REVIEW_PROMPT` |
| `REVIEWER_CLI` | `codex` \| `claude` \| `cursor` \| `opencode` |
| `REVIEWER_CLI` | `codex` \| `claude` \| `cursor` \| `opencode` \| `pi` |
| `REVIEWER_MODEL` | Model name |
| `MAX_ROUNDS` | Default 10 |
| `SESSION_ID_VAR` | `CODEX_PLAN_SESSION_ID` \| `CODEX_IMPL_SESSION_ID` \| `CURSOR_PLAN_SESSION_ID` \| `CURSOR_IMPL_SESSION_ID` \| `OPENCODE_PLAN_SESSION_ID` \| `OPENCODE_IMPL_SESSION_ID` |
@@ -456,6 +469,18 @@ Write the reviewer invocation to `/tmp/do-task-${REVIEW_KIND}-review-${REVIEW_ID
set -euo pipefail
```
**If `REVIEWER_CLI` is `pi`:**
Fresh call every round (Pi reviewer calls do not use session resume):
```bash
pi --no-session --no-skills --no-prompt-templates --no-extensions --no-context-files \
--model "$REVIEWER_MODEL" \
--tools read,grep,find,ls \
-p "Read the file /tmp/do-task-${REVIEW_KIND}-${REVIEW_ID}.md and review. Return exactly the required ## Summary, ## Findings, and ## Verdict structure."
```
**If `REVIEWER_CLI` is `codex`:**
Round 1 — fresh `codex exec`:
@@ -652,7 +677,7 @@ After the command completes:
cp /tmp/do-task-${REVIEW_KIND}-review-${REVIEW_ID}.runner.out \
/tmp/do-task-${REVIEW_KIND}-review-${REVIEW_ID}.md
```
- `claude`: promote `.runner.out` into the `.md` file:
- `claude` or `pi`: promote `.runner.out` into the `.md` file:
```bash
cp /tmp/do-task-${REVIEW_KIND}-review-${REVIEW_ID}.runner.out \
/tmp/do-task-${REVIEW_KIND}-review-${REVIEW_ID}.md
skills/do-task/codex/templates/task-plan.md
+1 -1
View File
@@ -9,7 +9,7 @@
| Created | YYYY-MM-DD |
| Slug | YYYY-MM-DD-<slug> |
| Runtime | codex |
| Reviewer CLI | codex \| claude \| cursor \| opencode |
| Reviewer CLI | codex \| claude \| cursor \| opencode \| pi |
| Reviewer Model | <model> |
| MAX_ROUNDS | 10 |
| Branch Strategy | current-branch \| worktree |
skills/do-task/cursor/SKILL.md
+27 -2
View File
@@ -132,6 +132,19 @@ If the user has already specified a reviewer CLI and model (e.g., "do task X, re
Store `REVIEWER_CLI`, `REVIEWER_MODEL`, and `MAX_ROUNDS` for Phases 5 and 8.
Reviewer CLI: `codex`, `claude`, `cursor`, `opencode`, `pi`, or `skip`.
If `REVIEWER_CLI=pi`, verify the Pi reviewer binary before entering the review loop:
```bash
pi --version
```
For shorthand `pi/<pi-model-name>`, split only on the first slash when the prefix is exactly `pi`; store the complete remainder in `REVIEWER_MODEL`. Examples: `pi/claude-opus-4-7` -> `claude-opus-4-7`, `pi/anthropic/claude-opus-4-7` -> `anthropic/claude-opus-4-7`, and `pi/openrouter/anthropic/claude-opus-4-7` -> `openrouter/anthropic/claude-opus-4-7`.
When `REVIEWER_CLI=pi`, the reviewer model is configured independently from the model running this workflow. If the model/provider is unavailable, surface helper stderr/status and use `pi --list-models [search]` to inspect configured models.
### Phase 4: Initialize Plan Workspace
Cursor Agent CLI has no plan-mode concept; there is no plan-mode guard here.
@@ -366,7 +379,7 @@ This subroutine is invoked twice per `do-task` run: once in Phase 5 (`REVIEW_KIN
| `REVIEW_ID` | 8-char hex (from `uuidgen`); reused across rounds of the same loop |
| `PAYLOAD_PATH` | `/tmp/do-task-${REVIEW_KIND}-${REVIEW_ID}.md` |
| `PROMPT_TEMPLATE` | `PLAN_REVIEW_PROMPT` or `IMPL_REVIEW_PROMPT` |
| `REVIEWER_CLI` | `codex` \| `claude` \| `cursor` \| `opencode` |
| `REVIEWER_CLI` | `codex` \| `claude` \| `cursor` \| `opencode` \| `pi` |
| `REVIEWER_MODEL` | Model name |
| `MAX_ROUNDS` | Default 10 |
| `SESSION_ID_VAR` | `CODEX_PLAN_SESSION_ID` \| `CODEX_IMPL_SESSION_ID` \| `CURSOR_PLAN_SESSION_ID` \| `CURSOR_IMPL_SESSION_ID` \| `OPENCODE_PLAN_SESSION_ID` \| `OPENCODE_IMPL_SESSION_ID` |
@@ -463,6 +476,18 @@ Write the reviewer invocation to `/tmp/do-task-${REVIEW_KIND}-review-${REVIEW_ID
set -euo pipefail
```
**If `REVIEWER_CLI` is `pi`:**
Fresh call every round (Pi reviewer calls do not use session resume):
```bash
pi --no-session --no-skills --no-prompt-templates --no-extensions --no-context-files \
--model "$REVIEWER_MODEL" \
--tools read,grep,find,ls \
-p "Read the file /tmp/do-task-${REVIEW_KIND}-${REVIEW_ID}.md and review. Return exactly the required ## Summary, ## Findings, and ## Verdict structure."
```
**If `REVIEWER_CLI` is `codex`:**
Round 1 — fresh `codex exec`:
@@ -659,7 +684,7 @@ After the command completes:
cp /tmp/do-task-${REVIEW_KIND}-review-${REVIEW_ID}.runner.out \
/tmp/do-task-${REVIEW_KIND}-review-${REVIEW_ID}.md
```
- `claude`: promote `.runner.out` into the `.md` file:
- `claude` or `pi`: promote `.runner.out` into the `.md` file:
```bash
cp /tmp/do-task-${REVIEW_KIND}-review-${REVIEW_ID}.runner.out \
/tmp/do-task-${REVIEW_KIND}-review-${REVIEW_ID}.md
skills/do-task/cursor/templates/task-plan.md
+1 -1
View File
@@ -9,7 +9,7 @@
| Created | YYYY-MM-DD |
| Slug | YYYY-MM-DD-<slug> |
| Runtime | cursor |
| Reviewer CLI | codex \| claude \| cursor \| opencode |
| Reviewer CLI | codex \| claude \| cursor \| opencode \| pi |
| Reviewer Model | <model> |
| MAX_ROUNDS | 10 |
| Branch Strategy | current-branch \| worktree |
skills/do-task/opencode/SKILL.md
+27 -2
View File
@@ -132,6 +132,19 @@ If the user has already specified a reviewer CLI and model (e.g., "do task X, re
Store `REVIEWER_CLI`, `REVIEWER_MODEL`, and `MAX_ROUNDS` for Phases 5 and 8.
Reviewer CLI: `codex`, `claude`, `cursor`, `opencode`, `pi`, or `skip`.
If `REVIEWER_CLI=pi`, verify the Pi reviewer binary before entering the review loop:
```bash
pi --version
```
For shorthand `pi/<pi-model-name>`, split only on the first slash when the prefix is exactly `pi`; store the complete remainder in `REVIEWER_MODEL`. Examples: `pi/claude-opus-4-7` -> `claude-opus-4-7`, `pi/anthropic/claude-opus-4-7` -> `anthropic/claude-opus-4-7`, and `pi/openrouter/anthropic/claude-opus-4-7` -> `openrouter/anthropic/claude-opus-4-7`.
When `REVIEWER_CLI=pi`, the reviewer model is configured independently from the model running this workflow. If the model/provider is unavailable, surface helper stderr/status and use `pi --list-models [search]` to inspect configured models.
### Phase 4: Initialize Plan Workspace
OpenCode has no plan-mode concept; there is no plan-mode guard here.
@@ -362,7 +375,7 @@ This subroutine is invoked twice per `do-task` run: once in Phase 5 (`REVIEW_KIN
| `REVIEW_ID` | 8-char hex (from `uuidgen`); reused across rounds of the same loop |
| `PAYLOAD_PATH` | `/tmp/do-task-${REVIEW_KIND}-${REVIEW_ID}.md` |
| `PROMPT_TEMPLATE` | `PLAN_REVIEW_PROMPT` or `IMPL_REVIEW_PROMPT` |
| `REVIEWER_CLI` | `codex` \| `claude` \| `cursor` \| `opencode` |
| `REVIEWER_CLI` | `codex` \| `claude` \| `cursor` \| `opencode` \| `pi` |
| `REVIEWER_MODEL` | Model name |
| `MAX_ROUNDS` | Default 10 |
| `SESSION_ID_VAR` | `CODEX_PLAN_SESSION_ID` \| `CODEX_IMPL_SESSION_ID` \| `CURSOR_PLAN_SESSION_ID` \| `CURSOR_IMPL_SESSION_ID` \| `OPENCODE_PLAN_SESSION_ID` \| `OPENCODE_IMPL_SESSION_ID` |
@@ -455,6 +468,18 @@ Write the reviewer invocation to `/tmp/do-task-${REVIEW_KIND}-review-${REVIEW_ID
set -euo pipefail
```
**If `REVIEWER_CLI` is `pi`:**
Fresh call every round (Pi reviewer calls do not use session resume):
```bash
pi --no-session --no-skills --no-prompt-templates --no-extensions --no-context-files \
--model "$REVIEWER_MODEL" \
--tools read,grep,find,ls \
-p "Read the file /tmp/do-task-${REVIEW_KIND}-${REVIEW_ID}.md and review. Return exactly the required ## Summary, ## Findings, and ## Verdict structure."
```
**If `REVIEWER_CLI` is `codex`:**
Round 1 — fresh `codex exec`:
@@ -651,7 +676,7 @@ After the command completes:
cp /tmp/do-task-${REVIEW_KIND}-review-${REVIEW_ID}.runner.out \
/tmp/do-task-${REVIEW_KIND}-review-${REVIEW_ID}.md
```
- `claude`: promote `.runner.out` into the `.md` file:
- `claude` or `pi`: promote `.runner.out` into the `.md` file:
```bash
cp /tmp/do-task-${REVIEW_KIND}-review-${REVIEW_ID}.runner.out \
/tmp/do-task-${REVIEW_KIND}-review-${REVIEW_ID}.md
skills/do-task/opencode/templates/task-plan.md
+1 -1
View File
@@ -9,7 +9,7 @@
| Created | YYYY-MM-DD |
| Slug | YYYY-MM-DD-<slug> |
| Runtime | opencode |
| Reviewer CLI | codex \| claude \| cursor \| opencode |
| Reviewer CLI | codex \| claude \| cursor \| opencode \| pi |
| Reviewer Model | <model> |
| MAX_ROUNDS | 10 |
| Branch Strategy | current-branch \| worktree |
skills/do-task/pi/SKILL.md
+8
View File
@@ -103,6 +103,14 @@ Reviewer CLI: `codex`, `claude`, `cursor`, `opencode`, `pi`, or `skip`
Store `REVIEWER_CLI`, `REVIEWER_MODEL`, and `MAX_ROUNDS`.
If `REVIEWER_CLI=pi`, verify the Pi reviewer binary before entering the review loop:
```bash
pi --version
```
For shorthand `pi/<pi-model-name>`, split only on the first slash when the prefix is exactly `pi`; store the complete remainder in `REVIEWER_MODEL`. Examples: `pi/claude-opus-4-7` -> `claude-opus-4-7`, `pi/anthropic/claude-opus-4-7` -> `anthropic/claude-opus-4-7`, and `pi/openrouter/anthropic/claude-opus-4-7` -> `openrouter/anthropic/claude-opus-4-7`.
When `REVIEWER_CLI=pi`, the reviewer model is configured independently from the pi model running this workflow. Use any configured pi model string, including provider-qualified model IDs. If the reviewer model or provider is unavailable, surface the review helper stderr/status and ask for a configured model; use `pi --list-models [search]` to inspect configured models.
The pi reviewer command rendered into `/tmp/do-task-${REVIEW_KIND}-review-${REVIEW_ID}.sh` must be isolated and read-only:
skills/implement-plan/claude-code/SKILL.md
+38 -1
View File
@@ -61,6 +61,19 @@ If the user has already specified a reviewer CLI and model (e.g., "implement the
Store `REVIEWER_CLI`, `REVIEWER_MODEL`, and `MAX_ROUNDS`. These values are fixed for the entire run.
Reviewer CLI: `codex`, `claude`, `cursor`, `opencode`, `pi`, or `skip`.
If `REVIEWER_CLI=pi`, verify the Pi reviewer binary before entering the review loop:
```bash
pi --version
```
For shorthand `pi/<pi-model-name>`, split only on the first slash when the prefix is exactly `pi`; store the complete remainder in `REVIEWER_MODEL`. Examples: `pi/claude-opus-4-7` -> `claude-opus-4-7`, `pi/anthropic/claude-opus-4-7` -> `anthropic/claude-opus-4-7`, and `pi/openrouter/anthropic/claude-opus-4-7` -> `openrouter/anthropic/claude-opus-4-7`.
When `REVIEWER_CLI=pi`, the reviewer model is configured independently from the model running this workflow. If the model/provider is unavailable, surface helper stderr/status and use `pi --list-models [search]` to inspect configured models.
### Phase 3: Set Up Worktree (REQUIRED SUB-SKILL)
Invoke `superpowers:using-git-worktrees` explicitly.
@@ -235,6 +248,18 @@ Write the reviewer invocation to `/tmp/milestone-review-${REVIEW_ID}.sh` as a ba
set -euo pipefail
```
**If `REVIEWER_CLI` is `pi`:**
Fresh call every round (Pi reviewer calls do not use session resume):
```bash
pi --no-session --no-skills --no-prompt-templates --no-extensions --no-context-files \
--model "$REVIEWER_MODEL" \
--tools read,grep,find,ls \
-p "Read the file /tmp/milestone-${REVIEW_ID}.md and review. Return exactly the required ## Summary, ## Findings, and ## Verdict structure."
```
**If `REVIEWER_CLI` is `codex`:**
```bash
@@ -372,7 +397,7 @@ jq -r '.result' /tmp/milestone-review-${REVIEW_ID}.json > /tmp/milestone-review-
```
- If `REVIEWER_CLI=codex`, extract `CODEX_SESSION_ID` from `/tmp/milestone-review-${REVIEW_ID}.runner.out` after the helper or fallback run. If the review text is only in `.runner.out`, move or copy the actual review body into `/tmp/milestone-review-${REVIEW_ID}.md` before verdict parsing.
- If `REVIEWER_CLI=claude`, promote stdout captured by the helper or fallback runner into the markdown review file:
- If `REVIEWER_CLI=claude` or `REVIEWER_CLI=pi`, promote stdout captured by the helper or fallback runner into the markdown review file:
```bash
cp /tmp/milestone-review-${REVIEW_ID}.runner.out /tmp/milestone-review-${REVIEW_ID}.md
@@ -424,6 +449,18 @@ If a revision contradicts the user's explicit requirements, skip it and note it
Rewrite `/tmp/milestone-review-${REVIEW_ID}.sh` for the next round. The script should contain the reviewer invocation only; do not run it directly.
**If `REVIEWER_CLI` is `pi`:**
Fresh call with prior-round context (Pi reviewer calls do not use session resume):
```bash
pi --no-session --no-skills --no-prompt-templates --no-extensions --no-context-files \
--model "$REVIEWER_MODEL" \
--tools read,grep,find,ls \
-p "You previously reviewed this milestone and requested revisions. Read the updated payload at /tmp/milestone-${REVIEW_ID}.md and re-review using the same ## Summary, ## Findings, and ## Verdict structure."
```
**If `REVIEWER_CLI` is `codex`:**
Resume the existing session:
skills/implement-plan/codex/SKILL.md
+38 -1
View File
@@ -94,6 +94,19 @@ If the user has already specified a reviewer CLI and model (e.g., "implement the
Store `REVIEWER_CLI`, `REVIEWER_MODEL`, and `MAX_ROUNDS`. These values are fixed for the entire run.
Reviewer CLI: `codex`, `claude`, `cursor`, `opencode`, `pi`, or `skip`.
If `REVIEWER_CLI=pi`, verify the Pi reviewer binary before entering the review loop:
```bash
pi --version
```
For shorthand `pi/<pi-model-name>`, split only on the first slash when the prefix is exactly `pi`; store the complete remainder in `REVIEWER_MODEL`. Examples: `pi/claude-opus-4-7` -> `claude-opus-4-7`, `pi/anthropic/claude-opus-4-7` -> `anthropic/claude-opus-4-7`, and `pi/openrouter/anthropic/claude-opus-4-7` -> `openrouter/anthropic/claude-opus-4-7`.
When `REVIEWER_CLI=pi`, the reviewer model is configured independently from the model running this workflow. If the model/provider is unavailable, surface helper stderr/status and use `pi --list-models [search]` to inspect configured models.
### Phase 3: Set Up Worktree (REQUIRED SUB-SKILL)
Invoke `superpowers:using-git-worktrees`.
@@ -268,6 +281,18 @@ Write the reviewer invocation to `/tmp/milestone-review-${REVIEW_ID}.sh` as a ba
set -euo pipefail
```
**If `REVIEWER_CLI` is `pi`:**
Fresh call every round (Pi reviewer calls do not use session resume):
```bash
pi --no-session --no-skills --no-prompt-templates --no-extensions --no-context-files \
--model "$REVIEWER_MODEL" \
--tools read,grep,find,ls \
-p "Read the file /tmp/milestone-${REVIEW_ID}.md and review. Return exactly the required ## Summary, ## Findings, and ## Verdict structure."
```
**If `REVIEWER_CLI` is `codex`:**
```bash
@@ -405,7 +430,7 @@ jq -r '.result' /tmp/milestone-review-${REVIEW_ID}.json > /tmp/milestone-review-
```
- If `REVIEWER_CLI=codex`, extract `CODEX_SESSION_ID` from `/tmp/milestone-review-${REVIEW_ID}.runner.out` after the helper or fallback run. If the review text is only in `.runner.out`, move or copy the actual review body into `/tmp/milestone-review-${REVIEW_ID}.md` before verdict parsing.
- If `REVIEWER_CLI=claude`, promote stdout captured by the helper or fallback runner into the markdown review file:
- If `REVIEWER_CLI=claude` or `REVIEWER_CLI=pi`, promote stdout captured by the helper or fallback runner into the markdown review file:
```bash
cp /tmp/milestone-review-${REVIEW_ID}.runner.out /tmp/milestone-review-${REVIEW_ID}.md
@@ -457,6 +482,18 @@ If a revision contradicts the user's explicit requirements, skip it and note it
Rewrite `/tmp/milestone-review-${REVIEW_ID}.sh` for the next round. The script should contain the reviewer invocation only; do not run it directly.
**If `REVIEWER_CLI` is `pi`:**
Fresh call with prior-round context (Pi reviewer calls do not use session resume):
```bash
pi --no-session --no-skills --no-prompt-templates --no-extensions --no-context-files \
--model "$REVIEWER_MODEL" \
--tools read,grep,find,ls \
-p "You previously reviewed this milestone and requested revisions. Read the updated payload at /tmp/milestone-${REVIEW_ID}.md and re-review using the same ## Summary, ## Findings, and ## Verdict structure."
```
**If `REVIEWER_CLI` is `codex`:**
Resume the existing session:
skills/implement-plan/cursor/SKILL.md
+38 -1
View File
@@ -94,6 +94,19 @@ If the user has already specified a reviewer CLI and model (e.g., "implement the
Store `REVIEWER_CLI`, `REVIEWER_MODEL`, and `MAX_ROUNDS`. These values are fixed for the entire run.
Reviewer CLI: `codex`, `claude`, `cursor`, `opencode`, `pi`, or `skip`.
If `REVIEWER_CLI=pi`, verify the Pi reviewer binary before entering the review loop:
```bash
pi --version
```
For shorthand `pi/<pi-model-name>`, split only on the first slash when the prefix is exactly `pi`; store the complete remainder in `REVIEWER_MODEL`. Examples: `pi/claude-opus-4-7` -> `claude-opus-4-7`, `pi/anthropic/claude-opus-4-7` -> `anthropic/claude-opus-4-7`, and `pi/openrouter/anthropic/claude-opus-4-7` -> `openrouter/anthropic/claude-opus-4-7`.
When `REVIEWER_CLI=pi`, the reviewer model is configured independently from the model running this workflow. If the model/provider is unavailable, surface helper stderr/status and use `pi --list-models [search]` to inspect configured models.
### Phase 3: Set Up Worktree (REQUIRED SUB-SKILL)
Invoke `superpowers:using-git-worktrees`.
@@ -272,6 +285,18 @@ Write the reviewer invocation to `/tmp/milestone-review-${REVIEW_ID}.sh` as a ba
set -euo pipefail
```
**If `REVIEWER_CLI` is `pi`:**
Fresh call every round (Pi reviewer calls do not use session resume):
```bash
pi --no-session --no-skills --no-prompt-templates --no-extensions --no-context-files \
--model "$REVIEWER_MODEL" \
--tools read,grep,find,ls \
-p "Read the file /tmp/milestone-${REVIEW_ID}.md and review. Return exactly the required ## Summary, ## Findings, and ## Verdict structure."
```
**If `REVIEWER_CLI` is `codex`:**
```bash
@@ -415,7 +440,7 @@ jq -r '.result' /tmp/milestone-review-${REVIEW_ID}.json > /tmp/milestone-review-
```
- If `REVIEWER_CLI=codex`, extract `CODEX_SESSION_ID` from `/tmp/milestone-review-${REVIEW_ID}.runner.out` after the helper or fallback run. If the review text is only in `.runner.out`, move or copy the actual review body into `/tmp/milestone-review-${REVIEW_ID}.md` before verdict parsing.
- If `REVIEWER_CLI=claude`, promote stdout captured by the helper or fallback runner into the markdown review file:
- If `REVIEWER_CLI=claude` or `REVIEWER_CLI=pi`, promote stdout captured by the helper or fallback runner into the markdown review file:
```bash
cp /tmp/milestone-review-${REVIEW_ID}.runner.out /tmp/milestone-review-${REVIEW_ID}.md
@@ -467,6 +492,18 @@ If a revision contradicts the user's explicit requirements, skip it and note it
Rewrite `/tmp/milestone-review-${REVIEW_ID}.sh` for the next round. The script should contain the reviewer invocation only; do not run it directly.
**If `REVIEWER_CLI` is `pi`:**
Fresh call with prior-round context (Pi reviewer calls do not use session resume):
```bash
pi --no-session --no-skills --no-prompt-templates --no-extensions --no-context-files \
--model "$REVIEWER_MODEL" \
--tools read,grep,find,ls \
-p "You previously reviewed this milestone and requested revisions. Read the updated payload at /tmp/milestone-${REVIEW_ID}.md and re-review using the same ## Summary, ## Findings, and ## Verdict structure."
```
**If `REVIEWER_CLI` is `codex`:**
Resume the existing session:
skills/implement-plan/opencode/SKILL.md
+38 -1
View File
@@ -76,6 +76,19 @@ If the user has already specified a reviewer CLI and model (e.g., "implement the
Store `REVIEWER_CLI`, `REVIEWER_MODEL`, and `MAX_ROUNDS`. These values are fixed for the entire run.
Reviewer CLI: `codex`, `claude`, `cursor`, `opencode`, `pi`, or `skip`.
If `REVIEWER_CLI=pi`, verify the Pi reviewer binary before entering the review loop:
```bash
pi --version
```
For shorthand `pi/<pi-model-name>`, split only on the first slash when the prefix is exactly `pi`; store the complete remainder in `REVIEWER_MODEL`. Examples: `pi/claude-opus-4-7` -> `claude-opus-4-7`, `pi/anthropic/claude-opus-4-7` -> `anthropic/claude-opus-4-7`, and `pi/openrouter/anthropic/claude-opus-4-7` -> `openrouter/anthropic/claude-opus-4-7`.
When `REVIEWER_CLI=pi`, the reviewer model is configured independently from the model running this workflow. If the model/provider is unavailable, surface helper stderr/status and use `pi --list-models [search]` to inspect configured models.
### Phase 4: Set Up Worktree (REQUIRED SUB-SKILL)
Use OpenCode's native skill tool to load:
@@ -253,6 +266,18 @@ Write the reviewer invocation to `/tmp/milestone-review-${REVIEW_ID}.sh` as a ba
set -euo pipefail
```
**If `REVIEWER_CLI` is `pi`:**
Fresh call every round (Pi reviewer calls do not use session resume):
```bash
pi --no-session --no-skills --no-prompt-templates --no-extensions --no-context-files \
--model "$REVIEWER_MODEL" \
--tools read,grep,find,ls \
-p "Read the file /tmp/milestone-${REVIEW_ID}.md and review. Return exactly the required ## Summary, ## Findings, and ## Verdict structure."
```
**If `REVIEWER_CLI` is `codex`:**
```bash
@@ -390,7 +415,7 @@ jq -r '.result' /tmp/milestone-review-${REVIEW_ID}.json > /tmp/milestone-review-
```
- If `REVIEWER_CLI=codex`, extract `CODEX_SESSION_ID` from `/tmp/milestone-review-${REVIEW_ID}.runner.out` after the helper or fallback run. If the review text is only in `.runner.out`, move or copy the actual review body into `/tmp/milestone-review-${REVIEW_ID}.md` before verdict parsing.
- If `REVIEWER_CLI=claude`, promote stdout captured by the helper or fallback runner into the markdown review file:
- If `REVIEWER_CLI=claude` or `REVIEWER_CLI=pi`, promote stdout captured by the helper or fallback runner into the markdown review file:
```bash
cp /tmp/milestone-review-${REVIEW_ID}.runner.out /tmp/milestone-review-${REVIEW_ID}.md
@@ -442,6 +467,18 @@ If a revision contradicts the user's explicit requirements, skip it and note it
Rewrite `/tmp/milestone-review-${REVIEW_ID}.sh` for the next round. The script should contain the reviewer invocation only; do not run it directly.
**If `REVIEWER_CLI` is `pi`:**
Fresh call with prior-round context (Pi reviewer calls do not use session resume):
```bash
pi --no-session --no-skills --no-prompt-templates --no-extensions --no-context-files \
--model "$REVIEWER_MODEL" \
--tools read,grep,find,ls \
-p "You previously reviewed this milestone and requested revisions. Read the updated payload at /tmp/milestone-${REVIEW_ID}.md and re-review using the same ## Summary, ## Findings, and ## Verdict structure."
```
**If `REVIEWER_CLI` is `codex`:**
Resume the existing session:
skills/implement-plan/pi/SKILL.md
+8
View File
@@ -86,6 +86,14 @@ Reviewer CLI: `codex`, `claude`, `cursor`, `opencode`, `pi`, or `skip`
Store `REVIEWER_CLI`, `REVIEWER_MODEL`, and `MAX_ROUNDS`.
If `REVIEWER_CLI=pi`, verify the Pi reviewer binary before entering the review loop:
```bash
pi --version
```
For shorthand `pi/<pi-model-name>`, split only on the first slash when the prefix is exactly `pi`; store the complete remainder in `REVIEWER_MODEL`. Examples: `pi/claude-opus-4-7` -> `claude-opus-4-7`, `pi/anthropic/claude-opus-4-7` -> `anthropic/claude-opus-4-7`, and `pi/openrouter/anthropic/claude-opus-4-7` -> `openrouter/anthropic/claude-opus-4-7`.
When `REVIEWER_CLI=pi`, the reviewer model is configured independently from the pi model running this workflow. Use any configured pi model string, including provider-qualified model IDs. If the reviewer model or provider is unavailable, surface the review helper stderr/status and ask for a configured model; use `pi --list-models [search]` to inspect configured models.
The pi reviewer command rendered into `/tmp/milestone-review-${REVIEW_ID}.sh` must be isolated and read-only: