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/adithyan-ak/agenthound/claude-mdgit clone --depth 1 https://github.com/adithyan-ak/AgentHoundWrote 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/adithyan-ak/agenthound/claude-md)<a href="https://agentmods.dev/instructions/adithyan-ak/agenthound/claude-md"><img src="https://agentmods.dev/badge/instructions/adithyan-ak/agenthound/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 5.1 | $0.01136 | $0.01136 |
| Opus 5 | $0.00568 | $0.00568 |
| Sonnet 5 | $0.00227 | $0.00227 |
| Haiku 4.5 | $0.00114 | $0.00114 |
Grade A, and why
AgentHound 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 5d 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 — 79 lines — stays where its author put it; the contents beside it link to each section on GitHub.
AgentHound maintainer contract
AgentHound provides autonomous offensive collection and access verification for AI agent infrastructure.
agenthoundis the static foothold-time collector. It scans, plans, verifies, restores, and checkpoints one local JSON artifact.agenthound-servermanually ingests artifacts, publishes the Neo4j graph and PostgreSQL finding state, and serves the API and dashboard.
Required checks
Before committing:
make check
Changes spanning collection, planning, artifacts, ingestion, or findings also run make integration. Documentation changes run make docs-check.
Before a release tag, run make prerelease and make docs-check. Release tags are numeric SemVer without a v prefix. The first numeric heading in CHANGELOG.md is the version source of truth; make sync-version updates every live installer, environment, and Docker Compose pin in the README, installer, and operator documentation.
Release from a clean commit already merged into main:
- Confirm every PR check is green and
Unreleasedis empty. - Dispatch the Release workflow from
mainonce to verify the repository-linked private staging packages. - Run
make prereleaseandmake docs-checkfrom the exactmaincommit. - Create and push an annotated numeric tag, for example
git tag -a 1.1.0 -m "AgentHound 1.1.0"followed bygit push origin 1.1.0. - Wait for the private candidate to pass, approve
release-production, and requirerelease-acceptanceandhomebrew-acceptanceto pass.
Numeric release tags are immutable. Rerun a transient failure against the same commit; if source changes are required, prepare the next patch version instead of moving the tag.
Supported release surfaces are GitHub archives, the installer, GHCR, and Homebrew. Numeric tags are intentionally not Go module versions, so release acceptance must not use go list, go mod download, or go install as distribution checks.
Hard boundaries
- The collector must not link
chi,pgx,neo4j-go-driver, orserver/internalpackages.scripts/deps-check.shenforces the boundary. - The stripped Linux amd64 collector must remain within the budget enforced by
scripts/size-check.sh. - Add every collector package and dependency to
scripts/collector-allowlist.txt. - Allowed dependency licenses are Apache-2.0, MIT, BSD-2-Clause, BSD-3-Clause, ISC, MPL-2.0, Unlicense, and Zlib.
- TLS verification is enabled by default. Every AgentHound-owned connection uses the shared contact policy through the final dial.
- The server is single-user and binds to loopback by default. OriginGuard protects browser mutations; callers without an
Originheader are inside the local-process trust boundary. - PostgreSQL and Neo4j form one storage pair and must be backed up and restored together.
- Build the UI before the server so
server/internal/api/ui/distcontains the files required bygo:embed.
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.
- 5d ago First seen · 79 lines · 1,136 tokens per session scan A 9296acda9633
AgentHound CLAUDE.md is an instructions file published in the GitHub repository adithyan-ak/AgentHound (382 stars, last pushed 2d ago), licensed Apache-2.0. It adds 1,136 tokens to every session, about $0.0057 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 2026-08-30.
Other instructions, from other repositories
Agentlas-OS AGENTS.md
AGENTS.md instructions for agentlas-ai/Agentlas-OS, covering agentlas core engine meta-agent team, repository constitution: local main only, public release allowlist (hard rule), source of truth and third-party plugin boundary.
Agentlas-OS GEMINI.md
Gemini CLI instructions for agentlas-ai/Agentlas-OS, covering gemini cli adapter, public release boundary, startup, default behavior and global command.
a2a-editor CLAUDE.md
Claude Code instructions for open-resource-discovery/a2a-editor, covering claude.md, project overview, commands, development and building.
ave CLAUDE.md
Claude Code instructions for aveproject/ave, covering claude.md — aveproject/ave, project, the critical distinction — read this twice, record schema v1.1.0 and adding a record.
MCP-Sentinel AGENTS.md
AGENTS.md instructions for BashaarJavaid/MCP-Sentinel, covering agents.md, project, where things live, conventions and commands.
hush CLAUDE.md
Instructions for backbay-labs/hush, covering claude.md, project overview, repo structure, common commands and rust.