pocmap CLAUDE.md

Project instructions for PocMap, a toolkit for finding and studying software vulnerabilities using a command-line tool, an MCP server, and a Python API. They also describe how agents should keep technical research current.

In plain words
What is it for?
Use it when developing or changing PocMap, especially for non-trivial integrations, tools, SDKs, platforms, or architecture. It points agents to the project's documentation and the separate guide for using its MCP tools.
Why use it?
It gives an agent the project's conventions, structure, commands, and research expectations in one place. This reduces decisions based on outdated knowledge or incorrect project assumptions.

Instructions file

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/zebbern/pocmap/claude-md
Clone the repo
git clone --depth 1 https://github.com/zebbern/pocmap
Per session 2,315 This file is loaded in full into every session.
When invoked 2,315 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 $0.02315 $0.02315
Opus 5 $0.01157 $0.01157
Sonnet 5 $0.00463 $0.00463
Haiku 4.5 $0.00231 $0.00231

Measured 2d ago against content hash ef344a33d0ee, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

pocmap CLAUDE.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 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.

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.

CLAUDE.md · 115 lines

How it starts

The opening of the file, as written. The whole thing — 115 lines — stays where its author put it; the contents beside it link to each section on GitHub.

PocMap Conventions

PocMap is an AI-friendly CVE / PoC / exploit-discovery toolkit (CLI + MCP server + Python API). This file covers how to develop the project. For how AI agents consume the MCP tools, see .claude/skills/pocmap-agent/references/mcp_tools.md (canonical MCP/agent guide; also .claude/skills/pocmap-agent/SKILL.md). AGENTS.md holds general agent guidelines, not the MCP tool contract. For full usage/API docs, see README.md.

Knowledge Freshness & Research Preferences (IMPORTANT!)

Do not rely solely on your internal knowledge or training data when making implementation decisions. Your memory is frequently outdated.

Before implementing anything non-trivial (especially integrations, tools, SDKs, platforms, AI agents, or architectural patterns):

  1. Actively look up the latest official documentation, release notes, changelogs, and recommended approaches.

  2. Check whether newer versions, official solutions, or better patterns already exist that would change how the task should be approached.

  3. Prefer using current best practices and existing tools over rebuilding something from scratch based on older knowledge.

  4. If there is any uncertainty about the current recommended way to do something, ask clarifying questions instead of defaulting to what you remember.

Why this matters:

Defaulting to internal knowledge often leads to outdated implementations. For example, an agent might implement an older pattern for something like MCP or an AI agent integration simply because that is what it “knows,” without checking whether a newer release, official SDK, or simpler approach has become available. Always verify first.

Commands

  • Install (PyPI): pip install "pocmap[server]" (omit [server] for CLI-only)
  • Install (dev clone): pip install -e ".[server,dev]" (pytest, mypy, ruff + the MCP SDK); docs site: pip install -e ".[server,docs]" then python scripts/generate_mcp_docs.py --check && mkdocs build --strict
  • Run CLI: pocmap lookup CVE-2021-44228 (or python -m pocmap ...); full command list: python -m pocmap --help
  • Lint: ruff check src/pocmap
  • Type check: mypy src (strict mode is on in pyproject.toml)
  • Run MCP server: uvx --from 'pocmap[server]' pocmap-mcp or installed pocmap-mcp / python -m pocmap.mcp_server (add --transport sse|http, --host, --port, --debug)
  • Tests: pytest works and is offline by default; see Testing below.

Read the full file on GitHub · 115 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 · 115 lines · 2,315 tokens per session scan A ef344a33d0ee

Subscribe to this mod's changes

pocmap CLAUDE.md is an instructions file published in the GitHub repository zebbern/pocmap (5 stars, last pushed 9d ago), licensed MIT. It adds 2,315 tokens to every session, about $0.0116 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.