Skip to content

docs: ground Copilot (and reviewers) on Garnet App comment - #119

Merged
jadoonf merged 1 commit into
mainfrom
cursor/garnet-reviewer-grounding-5b51
Oct 26, 2026
Merged

jadoonf merged 1 commit into
mainfrom
cursor/garnet-reviewer-grounding-5b51

Conversation

@jadoonf

@jadoonf jadoonf commented Oct 26, 2026 •

Copy link
Copy Markdown

Summary

Adds repository instructions that ground Copilot pull request reviews exclusively in the head-bound garnet-runtime-review[bot] App comment. Aligns the existing CodeRabbit and Qodo guidance on the exact runtime-evidence utterance prefix and safeguards while preserving their pnpm-specific role split, filters, and labels.

This is demo hygiene for Rank-2 reviewer consume. It changes no product behavior.

Squash Commit Body

Add Copilot pull request review instructions for consuming only head-bound Garnet App-comment evidence. Align CodeRabbit and Qodo on the same citation prefix and evidence safeguards without changing their existing review responsibilities.

This is reviewer demo hygiene and does not change pnpm product behavior.

Checklist

  • Updated the documentation if needed.

Written by an agent (Cursor, GPT-5.6 Sol).

Open in Web Open in Cursor 

View with [code]smith Autofix with [code]smith
Need help on this PR? Tag @codesmith-bot with what you need.

@coderabbitai

coderabbitai Bot commented Oct 26, 2026 •

Copy link
Copy Markdown

Review in Change Stack →

Caution

Review failed

The pull request is closed.

Finishing Touches
🧪 Generate unit tests (beta)
  • Commit to this branch
  • Create a new PR
  • Autofix · Keep fixing CodeRabbit findings and required CI, and resolving merge conflicts

Comment @coderabbitai help to get the list of available commands.

@jadoonf
jadoonf marked this pull request as ready for review October 26, 2026 18:04
Copilot AI balanced review requested due to automatic review settings October 26, 2026 18:04
@jadoonf
jadoonf merged commit 02e2bc2 into main Oct 26, 2026
26 of 28 checks passed
@qodo-code-review

Copy link
Copy Markdown

PR Summary by Qodo

Ground automated reviewers in head-bound Garnet comments

📝 Documentation ⚙️ Configuration changes 🕐 10-20 Minutes

Grey Divider

AI Description

• Add Copilot instructions to use only head-bound Garnet App comments as runtime evidence.
• Align CodeRabbit and Qodo on citation wording and safeguards while preserving their review roles.
• Clarify that missing or unfinished records cannot establish a clean run.
Diagram

graph TD
  A["Garnet App comment"] --> B{"Head SHA matches?"} -->|Yes| C["Evidence safeguards"] --> D["Copilot instructions"]
  C --> E["CodeRabbit guidance"] & F["Qodo guidelines"]
  B -->|No| G["Runtime undeterminable"]
Loading
High-Level Assessment

Keep the guidance in each reviewer's existing instruction surface. A separate shared policy file could reduce repetition, but would depend on every reviewer reliably loading it; the current approach makes the evidence rules explicit where each bot is configured.

Files changed (3) +35 / -2

Documentation (1) +33 / -0
copilot-instructions.mdAdd Copilot pull request review instructions +33/-0

Add Copilot pull request review instructions

• Directs Copilot to the repository review framework and defines when Garnet runtime evidence is usable. Specifies citations, treatment of missing records, and safeguards against treating recorded names or verdicts as authority.

.github/copilot-instructions.md

Other (2) +2 / -2
.coderabbit.yamlTighten CodeRabbit's Garnet evidence guidance +1/-1

Tighten CodeRabbit's Garnet evidence guidance

• Limits runtime evidence to the head-bound App comment and standardizes citation wording and safeguards. Leaves CodeRabbit's review focus and path-specific instructions intact.

.coderabbit.yaml

.pr_agent.tomlAlign Qodo's Garnet evidence guidance +1/-1

Align Qodo's Garnet evidence guidance

• Applies the same head-bound source restriction, citation prefix, and evidence safeguards to Qodo. Preserves its security- and performance-focused review responsibilities.

.pr_agent.toml

@qodo-code-review

Copy link
Copy Markdown

Code Review by Qodo

🐞 Bugs (2) 📘 Rule violations (0) 📜 Skill insights (0)

Grey Divider


Action required

1. Valid Garnet records are ignored 🐞 Bug ⛨ Security
Description
The new instructions recognize only garnet-runtime-review[bot], although the repository’s evidence
gate also trusts garnet-runtime-review-dev[bot] and garnet-ai[bot]. In environments where either
supported identity posts the head-bound Runtime Review record, reviewers following these
instructions can discard valid evidence and incorrectly describe runtime behavior as undeterminable.
Code

.github/copilot-instructions.md[R9-12]

+The only source of Garnet runtime evidence is the
+`garnet-runtime-review[bot]` pull request comment whose
+`<!-- garnet:commit <sha> -->` marker equals the pull request head SHA. Do not
+use Garnet evidence copied into the pull request description or any other
Relevance

●●● Strong

Documentation must match the gate’s trusted identities to avoid discarding valid head-bound runtime
evidence.

PR-#43
PR-#28

ⓘ Recommendations generated based on similar findings in past PRs

Evidence
The runtime evidence gate accepts all three identities, requires the Runtime Review marker, and
binds the record to the exact head SHA. The mirror uses the same trusted-author set, so the changed
documentation is inconsistent with the actual supported evidence sources.

.github/scripts/garnet-evidence-gate.mjs[20-29]
.github/scripts/garnet-evidence-gate.mjs[61-75]
.github/scripts/garnet-evidence-mirror.mjs[25-32]

Agent prompt
The issue below was found during a code review. Follow the provided context and guidance below and implement a solution

### Issue description
The reviewer instructions name only `garnet-runtime-review[bot]` as an authoritative Garnet commenter, but the repository’s evidence gate and mirror also trust `garnet-runtime-review-dev[bot]` and `garnet-ai[bot]`. Reviewers can therefore ignore valid head-bound records posted by those supported identities.

### Fix Focus Areas
- .github/copilot-instructions.md[9-12]
- .coderabbit.yaml[16-16]
- .pr_agent.toml[63-63]

### Recommended Fix
Describe the authoritative source as a Runtime Review comment from one of the identities accepted by the repository’s Garnet evidence gate, while retaining the exact head-commit binding and fail-closed rules.

ⓘ Copy this prompt and use it to remediate the issue with your preferred AI generation tools

Dismiss ↗ | View ↗


2. Unrelated bot comments look authoritative 🐞 Bug ⛨ Security
Description
The new instructions identify a Garnet record only by the bot identity and garnet:commit marker,
omitting the required garnet-runtime-review marker. A separate comment from a trusted Garnet
identity that carries the commit marker could consequently be treated by a reviewer as runtime
evidence even though the repository’s gate rejects it.
Code

.github/copilot-instructions.md[R9-12]

+The only source of Garnet runtime evidence is the
+`garnet-runtime-review[bot]` pull request comment whose
+`<!-- garnet:commit <sha> -->` marker equals the pull request head SHA. Do not
+use Garnet evidence copied into the pull request description or any other
Relevance

●●● Strong

Documentation should require the parser’s runtime-review marker, preventing unrelated trusted
comments from being treated as evidence.

PR-#43
PR-#28

ⓘ Recommendations generated based on similar findings in past PRs

Evidence
The changed instructions require only the commit marker, while the actual evidence parser first
requires the Runtime Review marker and then verifies the trusted author and exact head commit. This
mismatch can cause reviewers to accept a comment the repository’s own gate rejects.

.github/copilot-instructions.md[9-13]
.github/scripts/garnet-evidence-gate.mjs[20-29]
.github/scripts/garnet-evidence-gate.mjs[66-75]

Agent prompt
The issue below was found during a code review. Follow the provided context and guidance below and implement a solution

### Issue description
The documentation says that any comment from the named Garnet bot with a matching `garnet:commit` marker is authoritative, but the runtime evidence contract also requires `<!-- garnet-runtime-review -->`. Without that marker, reviewers may treat an unrelated trusted-bot comment as execution evidence.

### Fix Focus Areas
- .github/copilot-instructions.md[9-12]
- .coderabbit.yaml[16-16]
- .pr_agent.toml[63-63]

### Recommended Fix
State that the authoritative comment must be a trusted Garnet Runtime Review comment containing both `<!-- garnet-runtime-review -->` and a `garnet:commit` marker equal to the PR head SHA. Preserve the prohibition on using editable PR-description copies as authoritative evidence.

ⓘ Copy this prompt and use it to remediate the issue with your preferred AI generation tools

Dismiss ↗ | View ↗


Grey Divider

Context sources
✅ Compliance rules (platform): 64 rules
Review mode: Auto: 🚀 Fast: Localized reviewer-instruction documentation change with no product behavior impact.

Grey Divider

Tip of the day
💡 Did you know, you can type 'qodo, fix this' on a finding and the fix lands right on your PR

More tips ↗ | Customize Qodo ↗ | Qodo docs ↗

Grey Divider

Qodo Logo

Comment on lines +9 to +12
The only source of Garnet runtime evidence is the
`garnet-runtime-review[bot]` pull request comment whose
`<!-- garnet:commit <sha> -->` marker equals the pull request head SHA. Do not
use Garnet evidence copied into the pull request description or any other

Copy link
Copy Markdown

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Action required

1. Valid garnet records are ignored 🐞 Bug ⛨ Security

The new instructions recognize only garnet-runtime-review[bot], although the repository’s evidence
gate also trusts garnet-runtime-review-dev[bot] and garnet-ai[bot]. In environments where either
supported identity posts the head-bound Runtime Review record, reviewers following these
instructions can discard valid evidence and incorrectly describe runtime behavior as undeterminable.
Agent Prompt
### Issue description
The reviewer instructions name only `garnet-runtime-review[bot]` as an authoritative Garnet commenter, but the repository’s evidence gate and mirror also trust `garnet-runtime-review-dev[bot]` and `garnet-ai[bot]`. Reviewers can therefore ignore valid head-bound records posted by those supported identities.

### Fix Focus Areas
- .github/copilot-instructions.md[9-12]
- .coderabbit.yaml[16-16]
- .pr_agent.toml[63-63]

### Recommended Fix
Describe the authoritative source as a Runtime Review comment from one of the identities accepted by the repository’s Garnet evidence gate, while retaining the exact head-commit binding and fail-closed rules.

ⓘ Copy this prompt and use it to remediate the issue with your preferred AI generation tools

Dismiss ↗ | View ↗

Comment on lines +9 to +12
The only source of Garnet runtime evidence is the
`garnet-runtime-review[bot]` pull request comment whose
`<!-- garnet:commit <sha> -->` marker equals the pull request head SHA. Do not
use Garnet evidence copied into the pull request description or any other

Copy link
Copy Markdown

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Action required

2. Unrelated bot comments look authoritative 🐞 Bug ⛨ Security

The new instructions identify a Garnet record only by the bot identity and garnet:commit marker,
omitting the required garnet-runtime-review marker. A separate comment from a trusted Garnet
identity that carries the commit marker could consequently be treated by a reviewer as runtime
evidence even though the repository’s gate rejects it.
Agent Prompt
### Issue description
The documentation says that any comment from the named Garnet bot with a matching `garnet:commit` marker is authoritative, but the runtime evidence contract also requires `<!-- garnet-runtime-review -->`. Without that marker, reviewers may treat an unrelated trusted-bot comment as execution evidence.

### Fix Focus Areas
- .github/copilot-instructions.md[9-12]
- .coderabbit.yaml[16-16]
- .pr_agent.toml[63-63]

### Recommended Fix
State that the authoritative comment must be a trusted Garnet Runtime Review comment containing both `<!-- garnet-runtime-review -->` and a `garnet:commit` marker equal to the PR head SHA. Preserve the prohibition on using editable PR-description copies as authoritative evidence.

ⓘ Copy this prompt and use it to remediate the issue with your preferred AI generation tools

Dismiss ↗ | View ↗

@qodo-code-review

Copy link
Copy Markdown

PR Code Suggestions ✨

Warning

/improve is deprecated. Use /agentic_review instead (removal date not yet scheduled).

No code suggestions found for the PR.

Copilot AI left a comment

Copy link
Copy Markdown

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

🟢 Approval recommended

The instruction-only changes consistently implement the stated evidence policy, with no identified blocking issues.

No open issues 🎉

What changed in this PR

This PR aligns reviewer instructions for consuming Garnet runtime evidence without changing pnpm product behavior.

Changes:

  • Adds Copilot guidance restricting evidence to the Garnet App comment matching the PR head commit.
  • Aligns CodeRabbit and Qodo on citation formatting and safeguards against unsupported runtime claims.
  • Preserves existing reviewer responsibilities, filters, and labels.
File Description
.pr_agent.toml Aligns Qodo’s Garnet evidence and citation requirements.
.github/​copilot-instructions.md Adds Copilot review guidance and Garnet evidence safeguards.
.coderabbit.yaml Aligns CodeRabbit’s Garnet evidence and citation requirements.

Review effort: Balanced

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants