audit-instructions

A review of the instruction files that guide coding agents in a project, such as CLAUDE.md, AGENTS.md, rules, and skills. It compares those instructions with the project's actual files, history, tools, and tests.

In plain words
What is it for?
Use it to audit agent instructions, inspect the repository structure and recent history, and check whether the documented workflow matches the project's real setup.
Why use it?
It finds missing guidance, contradictions, hidden assumptions, and instructions that no longer fit the codebase. This helps agents work more reliably without guessing.

Command for Claude Code

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 commands/xsovad06/sova/audit-instructions
Clone the repo
git clone --depth 1 https://github.com/xsovad06/sova

Made for: Claude Code.

Per session 22 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 1,014 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 1 finding. 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 $0.00022 $0.01014
Opus 5 $0.00011 $0.00507
Sonnet 5 $0.00004 $0.00203
Haiku 4.5 $0.00002 $0.00101

Measured 2d ago against content hash 21269b1b5592, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

audit-instructions scanned grade A with 1 finding 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 2d 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.

Enumerates other installed skillslowAgent snooping

Other skills' SKILL.md files reveal prompts, capabilities and secrets that should be invisible to peers.

ls .claude/skills/*/SKILL.md 2>/dev/null || echo "No skills"

Downgraded: this mod is about security review, or the phrase is quoted, so it is likely naming the pattern rather than instructing it.

.claude/commands/audit-instructions.md · 120 lines

How it starts

The opening of the file, as written. The whole thing — 120 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Audit this project's AI instruction system for completeness, contradictions, implicit assumptions, and autonomous-agent readiness. $ARGUMENTS

Phase 1: Collect All Instruction Sources

Read every instruction file -- do not skip any.

cat CLAUDE.md 2>/dev/null || echo "MISSING: CLAUDE.md"
cat AGENTS.md 2>/dev/null || echo "No AGENTS.md"
for f in .claude/rules/*.md; do [ -f "$f" ] && echo "=== $f ===" && cat "$f"; done 2>/dev/null || echo "No rules"
ls .claude/agent-memory/*.md 2>/dev/null || echo "No agent memory"
ls .claude/skills/*/SKILL.md 2>/dev/null || echo "No skills"

Phase 2: Analyze Project Reality

Understand the actual project to compare against instructions:

find . -maxdepth 3 -type f \( -name "*.py" -o -name "*.ts" -o -name "*.js" -o -name "*.go" -o -name "*.rs" -o -name "*.java" \) | head -50
git log --oneline -20 2>/dev/null || echo "Not a git repo"
ls pyproject.toml package.json Cargo.toml go.mod Makefile Dockerfile 2>/dev/null
ls -d tests/ test/ __tests__/ spec/ 2>/dev/null && find tests/ test/ __tests__/ spec/ -maxdepth 1 -name "*.py" -o -name "*.ts" -o -name "*.js" 2>/dev/null | head -20

Phase 3: Gap Analysis

For each category, assess completeness (0-10) and list specific gaps:

Category What to check
Build commands Are test_cmd, lint_cmd, format_cmd documented and correct? Do they match actual tooling?
Architecture Does CLAUDE.md explain the module structure? Can an agent navigate the codebase without guessing?
Conventions Are naming conventions, file organization, and code style rules stated explicitly?
Error handling Are error handling patterns documented?
Testing Are test patterns, fixtures, mocks, and conventions documented?
Dependencies Are dependency management conventions documented (add/remove/upgrade)?
Security Are security-sensitive areas called out (credentials, external inputs, subprocess)?
Agent boundaries Are there clear "always do" / "ask first" / "never do" boundaries?
Contradictions Do any rules conflict with each other or with actual project practice?
Implicit assumptions What does the project assume that is not written down?

Read the full file on GitHub · 120 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. 2d ago First seen · 120 lines · 22 tokens per session scan A 21269b1b5592

Subscribe to this mod's changes

audit-instructions is a command published in the GitHub repository xsovad06/sova (2 stars, last pushed 2d ago), licensed Apache-2.0. It adds 22 tokens to every session and 1,014 once invoked, about $0.0001 per session on Opus 5. A static security scan graded it A with 1 finding (enumerates other installed skills). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.