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.
npx agentmods add commands/sembraniteam/claude-plugins/audit-fixgit clone --depth 1 https://github.com/sembraniteam/claude-pluginsWrote 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.
[](https://agentmods.dev/commands/sembraniteam/claude-plugins/audit-fix)<a href="https://agentmods.dev/commands/sembraniteam/claude-plugins/audit-fix"><img src="https://agentmods.dev/badge/commands/sembraniteam/claude-plugins/audit-fix.svg" alt="Measured on agentmods" height="20"></a>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.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5 | $0.00057 | $0.01098 |
| Opus 5 | $0.00028 | $0.00549 |
| Sonnet 5 | $0.00011 | $0.00220 |
| Haiku 4.5 | $0.00006 | $0.00110 |
Grade A, and why
audit-fix 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 3d 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.
How it starts
The opening of the file, as written. The whole thing — 101 lines — stays where its author put it; the contents beside it link to each section on GitHub.
You are running /audit-fix to remediate security findings with verified, minimal code changes.
Important: fixes are never applied automatically. This command always shows a plan before writing any file, and runs independent verification after.
Step 0 — Validate audit report exists
Before anything else, check whether an audit report is available:
- Does this session contain
/audit,/audit-diff, or/audit-fileoutput with a findings list? - Does a
SECURITY-AUDIT.mdfile exist in the current directory?
If neither is found, stop and say:
"No audit report found. Please run
/audit(or/audit-difffor a PR-scoped scan) first to generate a security findings list, then run/audit-fixagain."
Step 1 — Parse scope from $ARGUMENTS
Determine what to fix:
- Empty or
all→ all findings from the audit report - A single finding-ID (e.g.,
SA-001) → only that finding - A comma-separated list (e.g.,
SA-001,SA-003) → only those findings - A severity tier (e.g.,
critical,high) → all findings of that severity
Step 2 — Display scope and confirm with user
List the findings in scope, grouped by category:
**In scope for /audit-fix:**
Fixable findings (security-fixer will handle these):
SA-001 [Critical] SQL Injection in user lookup — src/db.js:14
SA-002 [High] Hardcoded API key — config.js:7
Potentially needs-human (fixer will confirm before skipping):
SA-003 [High] Auth session flow — may require architectural decision
Ask: "Should I proceed? The fixer will show a plan for each finding before writing anything. Reply 'yes' to confirm, or 'yes, fix all without asking per item' to skip per-finding plan confirmations."
Wait for the user's reply before proceeding.
Step 3 — Delegate to security-fixer
Spawn the security-fixer agent with:
- The audit report content (from this session or
SECURITY-AUDIT.md) - The finding-IDs in scope
- The user's batch approval preference (per-item confirmation or batch mode)
The fixer will work through findings one by one: show plan → wait for approval (unless batch) → apply change → move to next. Wait for the fixer to return its complete fix manifest before proceeding.
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.
- 3d ago First seen · 101 lines · 57 tokens per session scan A 03fa06c63189
audit-fix is a command published in the GitHub repository sembraniteam/claude-plugins (2 stars, last pushed 1mo ago), licensed MIT. It adds 57 tokens to every session and 1,098 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.
Other commands, from other repositories
git
Git operations with intelligent commit messages and workflow optimization.
checklist
Generate a custom checklist for the current feature based on user requirements.
clarify
Identify underspecified areas in the current feature spec by asking up to 5 highly targeted clarification questions and encoding answers back into the spec.
specify
Create or update the feature specification from a natural language feature description.
analyze
Perform a non-destructive cross-artifact consistency and quality analysis across spec.md, plan.md, and tasks.md after task generation.
converge
Assess the current codebase against the feature's spec, plan, and tasks, then append any remaining unbuilt work as new tasks to tasks.md so implement can complete it.