agentsclimarketplace

Athena receive review

Skill bencrooks-dev/Athena/skills/athena-receive-review

Handle incoming code review feedback with technical rigor — verify suggestions before implementing, push back on incorrect feedback, never agree performatively. Use when receiving PR comments, review notes, or correction requests.From its SKILL.md

Install
npx -y skills add bencrooks-dev/Athena --skill athena-receive-review

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

One thing to look at

  • 2 stars2 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

SKILL.md

6.4 KB, as published. Nobody here has run it

Athena Receive Review — Responder

Handle incoming code review feedback with intellectual honesty. Not every suggestion is correct. Not every criticism is valid. Your job is to evaluate feedback technically, not agree reflexively.

When This Fires

Invoke this skill when:

  • The user shares PR review comments or feedback
  • Someone (human or AI) has reviewed code and left suggestions
  • The user says "the reviewer said..." or "I got feedback that..."
  • You receive output from /athena-review or any code review tool
  • The user pastes review comments and asks you to address them

The Core Principle

Technical rigor over social compliance. Reviewers can be wrong. Agreeing with incorrect feedback and implementing bad suggestions is worse than pushing back respectfully. The codebase is the source of truth, not the reviewer's opinion.

Processing Feedback

Step 1: Catalog

List every piece of feedback as discrete items:

Review Feedback
═══════════════
1. [file:line] "suggestion or criticism"
2. [file:line] "suggestion or criticism"
3. [general] "suggestion or criticism"

Step 2: Triage Each Item

For EACH feedback item, classify it:

CategoryActionExample
Correct + ImportantImplement immediately"This SQL query is vulnerable to injection"
Correct + MinorImplement, low priority"Variable name could be clearer"
Subjective/StyleDiscuss with user, defer to project conventions"I'd prefer a different pattern here"
IncorrectPush back with evidence"This will cause a race condition" (when it won't)
UnclearAsk for clarification before acting"This doesn't look right" (what specifically?)

Step 3: Verify Before Implementing

For each item you're about to implement:

  1. Read the code the feedback references — does the criticism actually apply?
  2. Check the context — does the reviewer understand the surrounding code?
  3. Test the suggestion — would implementing it actually improve things?
  4. Look for side effects — will this change break something else?

Red flags that feedback might be wrong:

  • Reviewer references code that doesn't exist or has changed
  • Suggestion contradicts the project's established patterns
  • "Best practice" cited without considering the specific context
  • Reviewer didn't see related code that explains the choice
  • Suggestion would introduce a bug or regression

Step 4: Respond

For each item, report your assessment:

Feedback Response
═════════════════
1. [ACCEPT] file:line — Valid concern, implementing fix
2. [ACCEPT] file:line — Good catch, variable renamed
3. [DISCUSS] file:line — Subjective: reviewer prefers X, codebase uses Y. Keeping Y for consistency. [reason]
4. [REJECT] file:line — Incorrect: reviewer says this causes Z, but [evidence it doesn't]. Keeping as-is.
5. [CLARIFY] file:line — Need more detail: "doesn't look right" — what specifically?

Step 5: Implement Accepted Items

Only after triage and verification:

  1. Implement all ACCEPT items
  2. Run verification (invoke /athena-verify) to confirm changes don't break anything
  3. Report what was changed

Step 6: Log to Engram (if available)

If the Engram plugin is installed, call engram_log_session to record review feedback themes:

engram_log_session({
  topics: [
    "code review: <theme> (e.g., security, naming, error handling, testing)",
    "<specific pattern corrected> (e.g., SQL injection prevention, null checking)"
  ]
})

Focus on the category of feedback, not the specific file/line. This helps Engram detect recurring review themes — if "error handling" comes up in every review, that's a memory gap.

If feedback reveals a recurring correction (same mistake made before), suggest: "This correction has come up before. Consider saving it as an Engram feedback memory so it's caught during planning."

Handling Disagreements

When you believe feedback is incorrect:

  1. State your position clearly — "I disagree because [technical reason]"
  2. Show evidence — code, test output, documentation that supports your position
  3. Acknowledge uncertainty — "I could be wrong if [condition], but based on [evidence]..."
  4. Let the user decide — present both sides, recommend your position, let them call it

Never do this:

  • Agree with feedback you believe is wrong just to avoid conflict
  • Implement a suggestion that would introduce a bug
  • Say "good point" when it's not a good point
  • Rewrite working code because a reviewer has a style preference (unless it's a project convention)

Performative Agreement Detection

Watch for these patterns in YOUR OWN responses — they signal you're agreeing without thinking:

PatternProblem
"Great catch!" on every itemYou're not evaluating, you're performing
Implementing all feedback without questionSome feedback is always wrong or subjective
"You're right, I should have..."Maybe. Or maybe your original choice was fine.
Changing code without verifying the criticismBlind implementation
"Thanks for the thorough review" + implement allSocial compliance, not engineering

If you catch yourself doing this: STOP. Re-read the feedback. Evaluate each item independently.

Integration with Other Skills

  • After implementing accepted changes → invoke /athena-verify to confirm nothing broke
  • If feedback reveals a bug → invoke /athena-debug to investigate properly
  • If feedback requires significant rework → invoke /athena-plan to plan the changes

Rules

  • Verify before implementing — read the code, understand the context, then decide
  • Never agree performatively — intellectual honesty over social comfort
  • Evidence-based responses — "I disagree because [proof]" not "I disagree because I think so"
  • The codebase is truth — not the reviewer's memory of it
  • Accept valid feedback gracefully — being rigorous doesn't mean being defensive
  • Protect the user's code — implementing bad suggestions is worse than pushing back

Keep looking

Skills are one crate of 627,801. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.