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 instructions/dhevenddra/forensic-deepdive/claude-mdgit clone --depth 1 https://github.com/Dhevenddra/forensic-deepdiveWrote 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/instructions/dhevenddra/forensic-deepdive/claude-md)<a href="https://agentmods.dev/instructions/dhevenddra/forensic-deepdive/claude-md"><img src="https://agentmods.dev/badge/instructions/dhevenddra/forensic-deepdive/claude-md.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 | $0.01746 | $0.01746 |
| Opus 5 | $0.00873 | $0.00873 |
| Sonnet 5 | $0.00349 | $0.00349 |
| Haiku 4.5 | $0.00175 | $0.00175 |
Grade A, and why
forensic-deepdive 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 4d 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 — 91 lines — stays where its author put it; the contents beside it link to each section on GitHub.
CLAUDE.md — forensic-deepdive
What this project is
forensic-deepdive produces five durable markdown artifacts giving any AI coding agent forensic understanding of an unfamiliar codebase. Distillation over retention. Codified rules over vague summaries. Composable tools.
Five artifacts: MAP.md, HOTPATHS.md, ARCHAEOLOGY.md, MENTAL_MODEL.md, AGENT_BRIEF.md. AGENT_BRIEF is headline — ≤5kb, assertive Never/Always rules.
Stack
- Python 3.11+ (uv-managed);
typerCLI;rich(styled CLI, DEC-077);tree-sitter-language-pack;networkx;real-ladybug(graph, DEC-013);mcp;httpx;pydanticv2. - Optional extras:
graphiti(v0.2),semantic(ONNX, v0.3),openapi(v0.4),dev(ruff/pytest). - License: Apache-2.0.
Critical commands
uv sync --all-extras # install (dev tools = `dev` extra)
uv run forensic --version # smoke test
uv run forensic extract examples/tiny_fixture
uv run pytest tests/ -x # tests must pass before any merge
uv run ruff check src/ tests/
uv run ruff format src/ tests/
Session start protocol (MANDATORY)
- Read
CLAUDE.md(this file). - Read
DECISIONS.md— never contradict an active decision without writing a superseding entry first. - Read
PROGRESS.md— know what's done and what's next. - Glance at
git log --oneline -10to see the last 10 commits. - Confirm understanding in one sentence before writing any code: "Working on , respecting DEC-N about ."
Session end protocol (MANDATORY)
- Append to
PROGRESS.mdunder today's date: what was done, what's in flight, what's blocked. - If a non-trivial architectural choice was made, write a new
DECISIONS.mdentry. - Stage and commit with a conventional-commit message (
feat:,fix:,chore:,docs:,test:). - Never push without explicit user instruction.
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.
- 4d ago First seen · 91 lines · 1,746 tokens per session scan A 5a26181852b2
forensic-deepdive CLAUDE.md is an instructions file published in the GitHub repository Dhevenddra/forensic-deepdive (2 stars, last pushed 19d ago), licensed Apache-2.0. It adds 1,746 tokens to every session, about $0.0087 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 instructions, from other repositories
ChatCrystal AGENTS.md
AGENTS.md instructions for ZengLiangYi/ChatCrystal, covering agents.md, project overview, commands, development (server port 3721 + client port 13721) and build for production.
Dragon-Brain CLAUDE.md
Instructions for iikarus/Dragon-Brain, covering dragon brain — claude.md, the harness, audit remediation (april–may 2026, complete 2026-05-09), the lie this audit closed and the contract that matters now.
navegador CLAUDE.md
Instructions for ConflictHQ/navegador, covering navegador — claude context, what it is, stack, package layout and falkordb connection.
engram CLAUDE.md
Instructions for NickCirv/engram, covering codebase structure (auto-generated by engram), core entities, structure and key dependencies.
navegador AGENTS.md
Instructions for ConflictHQ/navegador, a project described as: AST + knowledge graph context engine for AI coding agents.
gossiphs AGENTS.md
Instructions for williamfzc/gossiphs, covering development habits and architectural principles, 1. metric priorities, 2. dependency extraction philosophy, 3. handling symbol fqns and 4. evaluation habits.