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/juwon1405/agentic-dart/agents-mdgit clone --depth 1 https://github.com/Juwon1405/agentic-dartWhat 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 | $0.01035 | $0.01035 |
| Opus 5 | $0.00517 | $0.00517 |
| Sonnet 5 | $0.00207 | $0.00207 |
| Haiku 4.5 | $0.00103 | $0.00103 |
Grade A, and why
agentic-dart AGENTS.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.
How it starts
The opening of the file, as written. The whole thing — 86 lines — stays where its author put it; the contents beside it link to each section on GitHub.
AGENTS.md — Agent Guide for Agentic-DART
Concise operational guide for Codex and other coding agents. For the full rationale see CLAUDE.md. When the two overlap, both are authoritative; CLAUDE.md carries the longer explanation.
No hard-coded counts live in this file on purpose — they drift on every ship and would violate our measure-don't-guess rule. Run the commands below for live values.
Project summary
Autonomous DFIR agent. A senior-analyst loop calls typed, read-only MCP tools (two layers: native pure-Python functions + SIFT-tool adapters), logs every call in a SHA-256-chained audit, and emits a findings report. Python 3.10+, MIT, SANS FIND EVIL! 2026.
Repository map
dart_audit/— SHA-256-chained JSONL audit logger (Python package)dart_mcp/— custom MCP server, typed read-only forensic functions (Python package)dart_agent/— iteration controller, deterministic + live loops, auth (Python package)dart_corr/— DuckDB cross-artifact correlation engine (Python package)dart_playbook/— senior-analyst YAML playbooks; data, nopyproject.toml, loaded by pathexamples/—case-studies/{self-evaluation,external-evaluation}/case-NN/(README + truth.json + evidence_root), demosscripts/—install.sh,benchmark/,scripts/eval/demo.py,generate_realistic_evidence.pytests/— main pytest suite;dart_corr/tests/— correlation-engine tests
Preferred commands
export PYTHONPATH=dart_audit/src:dart_mcp/src:dart_agent/src:dart_corr/src
# Full suite — every test must pass
python3 -m pytest tests/ dart_corr/tests/ -q
# Focused
python3 tests/test_mcp_surface.py # tool-surface drift (asserts the exact set)
python3 tests/test_mcp_bypass.py # adversarial / read-only guard
python3 -m pytest dart_corr/tests/ -q # correlation engine
# Live tool count / native / SIFT split
PYTHONPATH=dart_mcp/src python3 -c "import dart_mcp; t=dart_mcp._REGISTRY; \
s=[k for k in t if k.startswith('sift_')]; print(len(t), len(t)-len(s), len(s))"
# Offline demo (deterministic, no API key)
bash examples/demo-run.sh
# Accuracy (deterministic; recall must stay 1.0, hallucination 0)
python3 -m scripts.eval.demo
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 · 86 lines · 1,035 tokens per session scan A a7ddde1cd1a7
agentic-dart AGENTS.md is an instructions file published in the GitHub repository Juwon1405/agentic-dart (9 stars, last pushed 2mo ago), licensed MIT. It adds 1,035 tokens to every session, about $0.0052 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
ai-dfir-toolkit CLAUDE.md
Instructions for depalmar/ai-dfir-toolkit, covering project context, what this is, commands, rules that are not negotiable and what a restricted runner cannot verify.
SAVVYDFIR-MCP CLAUDE.md
Instructions for kismatkunwar89/SAVVYDFIR-MCP, covering savvydfir-mcp - dfir investigation framework, what this is, critical rules, forensic knowledge in tool responses and investigation workflow (7 phases).
sift-mcp AGENTS.md
Instructions for AppliedIR/sift-mcp, covering valhuntir — forensic investigation platform, getting started, available mcp servers, malware analysis escalation and investigation recording.
sentry-selfhosted-mcp AGENTS.md
Instructions for HieuAnh87/sentry-selfhosted-mcp, covering selfhosted-sentry-mcp, commands, install deps, build typescript to ./build and set executable bit and watch mode during source edits.
mem_forensics-mcp CLAUDE.md
Instructions for x746b/mem_forensics-mcp, covering role: memory forensics specialist, mem-forensics-mcp: triage first, architecture, investigation questions -> tools and workflow.
sentry-selfhosted-mcp CLAUDE.md
Instructions for HieuAnh87/sentry-selfhosted-mcp, covering selfhosted-sentry-mcp — claude code overrides, claude-specific and quick reminders.