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/tarekmo0/hunterstick/agents-mdgit clone --depth 1 https://github.com/tarekmo0/HunterStickWrote 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/tarekmo0/hunterstick/agents-md)<a href="https://agentmods.dev/instructions/tarekmo0/hunterstick/agents-md"><img src="https://agentmods.dev/badge/instructions/tarekmo0/hunterstick/agents-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 | $0.07183 | $0.07183 |
| Opus 5 | $0.03592 | $0.03592 |
| Sonnet 5 | $0.01437 | $0.01437 |
| Haiku 4.5 | $0.00718 | $0.00718 |
Grade A, and why
HunterStick AGENTS.md 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 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.
Makes network callslowCapability
Not a fault in itself. Listed so you know the mod talks to something, and to what.
| `recon-toolkit.md` · `mcp-tooling.md` | fallback only — the skill file names its tools · browser vs curl vs Burp | How it starts
The opening of the file, as written. The whole thing — 380 lines — stays where its author put it; the contents beside it link to each section on GitHub.
HunterStick — operating instructions
Read this file in full before doing anything. It is the canonical instruction set for this
repository. CLAUDE.md and the other runtime files are pointers to it.
Session start
Run this first. It is the same entry on every runtime:
tools/brief.py
It prints a compact snapshot of the named targets, their manifest state, and blockers. Add
--rules prints this file verbatim — there is no summarised second copy to drift. It reads the target files rather than maintaining another
state store. index.md holds decisions and state, not counts — ask kb.py runs for the
ledger and coverage.py for what is untested. Paste
the --rules output as your first message on a runtime that does not auto-load this file — or
just tell the model to read AGENTS.md, which is better whenever it can reach the filesystem.
Then read, in this order, only what applies:
| # | file | why |
|---|---|---|
| 0 | tools/config.py check |
machine config — researcher identity, VPS, API keys. Never read config.json directly; it holds secrets. check redacts credentials. Unset values are named blockers, not silent defaults. |
| 1 | targets/TARGETS.md when present |
the local, gitignored registry. tools/brief.py also discovers target directories directly. There is no "active" target. |
| 2 | README.md · TOOLS.md |
only when explaining or auditing HunterStick itself, not during an ordinary hunt. GitHub overview · tool reference. |
| 2a | playbooks/system/orchestrator.md |
always load, first. Who you are — a security researcher, your edge, what you are accountable for — then the loop, the delegation rules, and the boundary you do not cross. |
| 3 | playbooks/system/class-signals.md |
load when the task is orient / suggest / "what next?" — request analysis, property derivation, coverage, or recommending a class/phase/flow the human did not ask for. It is the catalogue: surface-property vocabulary → 45 vuln classes, signal → the recon skills, situation → 10 flows, one line each. You cannot recommend what you do not know exists — but for an exactly named skill you do not need the whole catalogue, so skip it and save the context. A map, never a substitute for the file it points at. |
| 4 | targets/<named>/index.md |
the manifest of the target the human named — nothing else |
| 5 | targets/<named>/target.md · scope.md |
profile · what the hunter said they are testing. scope.md always loads because it is context you reason with — a model that enumerates can widen one apex into three thousand names. It is ~30 lines. It is not a gate: see Scope is the hunter's below. |
| 5a | tools/app-context.py resolve targets/<named> <host-or-url> when the task names one exact application |
prints only that app's map.md, endpoints.md, and notes.md. Load those files if present. A target-wide or multi-host task does not select one app, so it does not load all sibling applications. |
| 6 | accounts.md only when the task involves a principal |
authenticated recon, authz, IDOR, multi-tenant, privesc. Unauthenticated recon does not need it and should not pay for it. |
| — | scope-evidence.md only on a dispute |
where the scope statement came from. You consult provenance when something disagrees, not every turn. |
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 · 380 lines · 7,183 tokens per session scan A 3cd9aee211fb
HunterStick AGENTS.md is an instructions file published in the GitHub repository tarekmo0/HunterStick (13 stars, last pushed 17d ago), licensed MIT. It adds 7,183 tokens to every session, about $0.0359 per session on Opus 5. A static security scan graded it A with 1 finding (makes network calls). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other instructions, from other repositories
vscode buildNext.instructions.md
Working notes and architecture documentation for the new esbuild-based build system in build/next. Use when making changes to the new build pipeline (transpile/bundle commands, NLS plugin, source-map handling, resource copying, or self-hosting watch tasks).
spec-kit AGENTS.md
AGENTS.md instructions for github/spec-kit, covering agents.md, about spec kit and specify, quickstart — add a new integration in 5 steps, integration architecture and integrationmanifest — file tracking.
codex AGENTS.md
AGENTS.md instructions for openai/codex, covering rust/codex-rs, the codex-core crate, code review rules, crate api surface and model visible context.
langchain AGENTS.md
AGENTS.md instructions for langchain-ai/langchain, covering global development guidelines for the langchain monorepo, corridor security analysis, project architecture and context, monorepo structure and development tools & commands.
vscode oss-third-party-notices.instructions.md
Instructions for microsoft/vscode, covering vs code oss third-party-notices pipeline, architecture, pipeline flow in ci, applying the notice (cutover) and fallback chain (never fail the build).
next.js AGENTS.md
Instructions for vercel/next.js, covering next.js development guide, codebase structure, monorepo overview, core package: packages/next and other important packages.