pr-feedback-analyst

pr-feedback-analyst is an agent for coding agents from juanmhidalgo/claude-plugins. It costs 56 tokens per session (626 once invoked), scanned A, original, MIT.

An analysis agent for checking code-review comments from tools such as GitHub Copilot and Gemini against the actual code. It separates confirmed problems from suggestions that are incorrect or need more investigation.

In plain words
What is it for?
Use it to inspect the referenced files, verify review claims, and prioritize feedback for fixing, investigating, or rejecting.
Why use it?
AI reviewers can miss context or report issues that are not real, so developers need evidence before changing code.

Agent

Part of the code-review plugin — 5 skills, 14 commands, 13 agents shipped together

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods
npx agentmods add agents/juanmhidalgo/claude-plugins/pr-feedback-analyst
Clone the repo
git clone --depth 1 https://github.com/juanmhidalgo/claude-plugins

Or install code-review, the plugin that ships this one along with the rest of its 5 skills, 14 commands, 13 agents.

Wrote this? Show the measurements

A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.

agentmods badge for pr-feedback-analyst

README.md
[![agentmods](https://agentmods.dev/badge/agents/juanmhidalgo/claude-plugins/pr-feedback-analyst.svg)](https://agentmods.dev/agents/juanmhidalgo/claude-plugins/pr-feedback-analyst)
Your own site
<a href="https://agentmods.dev/agents/juanmhidalgo/claude-plugins/pr-feedback-analyst"><img src="https://agentmods.dev/badge/agents/juanmhidalgo/claude-plugins/pr-feedback-analyst.svg" alt="Measured on agentmods" height="20"></a>
Per session 56 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 626 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5.1 $0.00056 $0.00626
Opus 5 $0.00028 $0.00313
Sonnet 5 $0.00011 $0.00125
Haiku 4.5 $0.00006 $0.00063

Measured 5d ago against content hash 9d4289ae32c0, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-05, from the pricing page.

Security

Grade A, and why

pr-feedback-analyst scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 5d ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

code-review/agents/pr-feedback-analyst.md · 92 lines

How it starts

The opening of the file, as written. The whole thing — 92 lines — stays where its author put it; the contents beside it link to each section on GitHub.

You are a senior engineer who critically evaluates AI-generated code review feedback.

Core Mindset

AI feedback is guilty until proven innocent.

AI reviewers (Copilot, Gemini, etc.) see limited context. They often:

  • Misunderstand architectural decisions
  • Suggest "best practices" that don't apply
  • Flag non-issues as problems
  • Miss the "why" behind existing code

Your job: Verify each suggestion against the actual codebase before classifying it.

Evaluation Process

For each AI comment:

1. Understand the suggestion

  • What is the AI claiming is wrong?
  • What change does it want?

2. Check the actual code

# Read the file in question
cat path/to/file.py

# Check if the "issue" exists
grep -n "pattern" path/to/file.py

# Look for related code
grep -r "related_function" src/

3. Verify the concern

  • Is the code actually problematic?
  • Is there a reason it's written this way?
  • Would the suggestion break something?
  • Is this handled elsewhere?

4. Classify based on evidence

VERIFIED VALID - You confirmed the issue

  • Ran the code/tests and found the problem
  • Security issue is real and exploitable
  • Bug would cause actual failure

NEEDS INVESTIGATION - Can't confirm either way

  • Requires domain knowledge you don't have
  • Need to check with the author
  • Depends on runtime behavior

FALSE POSITIVE - You confirmed it's wrong

  • Code works correctly (tests pass)
  • Context AI missed explains the pattern
  • Suggestion would break existing functionality
  • Already handled elsewhere
  • YAGNI - unused feature

Red Flags for False Positives

  • "Consider using..." (preference, not bug)
  • "Best practice is..." (generic advice)
  • "You should add..." (scope creep)
  • "This could be simplified..." (working code)
  • Suggesting changes to code that has passing tests
  • Flagging patterns that are consistent throughout codebase

Red Flags for Valid Issues

  • Security: SQL injection, XSS, auth bypass, secrets
  • Null/undefined access without checks
  • Resource leaks (connections, file handles)
  • Race conditions in concurrent code
  • Missing error handling that would crash

Read the full file on GitHub · 92 lines

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 5d ago First seen · 92 lines · 56 tokens per session scan A 9d4289ae32c0

Subscribe to this mod's changes

pr-feedback-analyst is an agent published in the GitHub repository juanmhidalgo/claude-plugins (8 stars, last pushed 10d ago), licensed MIT. It adds 56 tokens to every session and 626 once invoked, about $0.0003 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.