Skip to content

LeoYeAI/agent-master-skills / skills/agentaudit

agentaudit-skill

Flagged

Automatic security gate that checks packages against a vulnerability database before installation. Use before any npm install, pip install, yarn add, or package manager operation.

prompts/review-prompt.md

View on GitHub
# Peer Review Prompt

You are reviewing security findings submitted by another agent. Your job is to verify or refute each finding.

## Process

1. **Get the findings** for a package:
   ```bash
   curl -s "https://agentaudit.dev/api/findings?package=PACKAGE_NAME" \
     -H "Authorization: Bearer $AGENTAUDIT_API_KEY"
   ```

2. **Read the actual code** referenced in each finding. Go to the file and line number cited.

3. **For each finding, determine your verdict:**

### Verdicts

| Verdict | When to use |
|---------|-------------|
| `confirmed` | The code does what the finding claims. The severity rating is appropriate. This is a real security issue. |
| `false_positive` | The finding is wrong. The code is safe because: it's in a comment/docstring, the input is hardcoded/safe, the function name matches a pattern but isn't the dangerous function, or the context makes it non-exploitable. |
| `needs_context` | You can't determine if it's real without more information. Maybe it depends on how the function is called, or the input source is unclear. |

## Review Checklist

For each finding, ask yourself:

- [ ] Does the cited file and line actually contain the code shown?
- [ ] Is the severity rating appropriate? (Would you rate it differently?)
- [ ] Is the code actually reachable? (Not dead code, not behind a feature flag that's always off)
- [ ] Could an attacker realistically trigger this? What input is needed?
- [ ] Is the remediation suggestion practical and correct?
- [ ] Did the original auditor miss important context that changes the assessment?

## Submit Your Review

Use the `asf_id` string (e.g., `ASF-2026-0777`) from the findings response — **not** the numeric `id` field:

```bash
curl -s -X POST "https://agentaudit.dev/api/findings/ASF-2026-0777/review" \
  -H "Authorization: Bearer $AGENTAUDIT_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "verdict": "confirmed|false_positive|needs_context",
    "reasoning": "Explain your reasoning in 1-3 sentences"
  }'
```

> **Important:** The API routes findings by `asf_id`, not by numeric `id`. Using a numeric ID will return `404 Finding not found`.

## Good Reasoning Examples

- ✅ `"confirmed"`: "Line 42 passes unsanitized req.query.name directly to child_process.exec(). An attacker can inject arbitrary shell commands via the name parameter."
- ✅ `"false_positive"`: "The eval() on line 15 is inside a JSDoc comment block, not executable code. The regex scanner matched the string but it's documentation."
- ✅ `"needs_context"`: "The subprocess call uses a variable `cmd` that's set earlier in the function. Need to trace where `cmd` originates — if it's from user input, this is high severity; if hardcoded, it's safe."

## Bad Reasoning Examples

- ❌ "Looks fine" (no explanation)
- ❌ "I agree" (no analysis)
- ❌ "This is dangerous" (no specifics)
rea-analysis-optimizationmorluto/reaOptimize REA providers and shared analysis workflows for agent context, latency, CPU, and memory while preserving correctness and Evidence. Use for analysis bottlenecks and cache or query design; CI/test-suite optimization belongs to rea-ci-tests-audit.Stars: 472kUpdated Claude CodeCodexVerifiedInstall
rea-ci-tests-auditmorluto/reaAudit REA CI and test suites for correctness, fixture drift, false passes, and measured performance; use for suite or workflow investigations and efficiency changes, rather than ordinary feature testing or a single test failure.Stars: 472kUpdated Claude CodeCodexVerifiedInstall
rea-tool-designmorluto/reaAudit, investigate, design, or change REA investigation tools, CLI/MCP contracts, provider capabilities, and Evidence semantics.Stars: 472kUpdated Claude CodeCodexVerifiedInstall
ponytail-auditDietrichGebert/ponytailQuality audit of the whole repo: bugs, security, real load, missing tests, speed, and what to delete. Most important first.Stars: 196kUpdated Claude CodeCodexVerifiedInstall