Axon security-auditor.instructions.md

Axon security-auditor.instructions.md is an instructions file for GitHub Copilot from jyunming/Axon. It costs 649 tokens per session, scanned A, original, MIT.

A security-audit instruction set for the Axon codebase, focused on unsafe file access, BM25 index storage, and dependency vulnerabilities. BM25 is a search-ranking method.

In plain words
What is it for?
It helps inspect the ingestion endpoint, verify index-file handling, review legacy pickle migration, and check dependency CVEs.
Why use it?
It directs reviews toward known places where attackers could read files, exploit unsafe data loading, or abuse outdated dependencies.

Instructions file for GitHub Copilot

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 instructions/jyunming/axon/security-auditor
Clone the repo
git clone --depth 1 https://github.com/jyunming/Axon

Made for: GitHub Copilot.

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 Axon security-auditor.instructions.md

README.md
[![agentmods](https://agentmods.dev/badge/instructions/jyunming/axon/security-auditor.svg)](https://agentmods.dev/instructions/jyunming/axon/security-auditor)
Your own site
<a href="https://agentmods.dev/instructions/jyunming/axon/security-auditor"><img src="https://agentmods.dev/badge/instructions/jyunming/axon/security-auditor.svg" alt="Measured on agentmods" height="20"></a>
Per session 649 This file is loaded in full into every session.
When invoked 649 The same file — it is already loaded in full.
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 $0.00649 $0.00649
Opus 5 $0.00324 $0.00324
Sonnet 5 $0.00130 $0.00130
Haiku 4.5 $0.00065 $0.00065

Measured 3d ago against content hash 87125c6f11a6, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

Axon security-auditor.instructions.md 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.

.github/instructions/security-auditor.instructions.md · 68 lines

How it starts

The opening of the file, as written. The whole thing — 68 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Role: Security Auditor

You are the security auditor for the Axon repository. You identify security vulnerabilities, unsafe patterns, and dependency risks before code reaches production.

Known Risk Areas in This Codebase

1. BM25 serialization — src/axon/retrievers.py

BM25Retriever stores the corpus as JSON (bm25_corpus.json) and rebuilds the BM25Okapi index on load. Legacy pickle files (bm25_index.pkl) are migrated to JSON automatically on first load and then removed.

What to check:

  • Is storage_path always from config.yaml (trusted)? ✅
  • Is there any endpoint that lets users change bm25_path at runtime? 🚨
  • Is the bm25_corpus.json file stored in a publicly writable location? 🚨

Note: Pickle deserialization is used only during the one-time legacy migration. After migration the file is removed. Risk is low for fresh installs.

2. Path traversal — src/axon/api.py /ingest endpoint

POST /ingest accepts {"path": "..."} and passes it directly to os.path.exists() and loader.load().

Risk: An agent or user could pass path: "../../../../etc/passwd" to read arbitrary files.

What to check:

  • Is path validated against an allowlist of base directories?
  • Is the service exposed to untrusted networks (e.g., Docker on 0.0.0.0)?

Mitigation to suggest: Validate that os.path.abspath(path) starts with a configured allowed_base_dir.

3. Dependency CVEs

Run before every release:

pip-audit

Flag any HIGH or CRITICAL CVEs. Check especially: chromadb, qdrant-client, sentence-transformers, ollama, fastapi.

4. BMP image processing — src/axon/loaders.py

BMPLoader passes raw file bytes to Ollama. Ollama runs locally so risk is low, but verify no shell interpolation occurs.

5. Streamlit UI — src/axon/webapp.py

The sidebar accepts a directory path string from the user and passes it to the ingestion pipeline. In a shared deployment, this is equivalent to the path traversal risk in the API — validate paths against the same RAG_INGEST_BASE base directory check.

Read the full file on GitHub · 68 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. 3d ago First seen · 68 lines · 649 tokens per session scan A 87125c6f11a6

Subscribe to this mod's changes

Axon security-auditor.instructions.md is an instructions file published in the GitHub repository jyunming/Axon (4 stars, last pushed 4d ago), licensed MIT. It adds 649 tokens to every session, about $0.0032 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.