Borrowing it
Nothing to install: this file belongs to IngTian/witness. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.
curl -O https://raw.githubusercontent.com/IngTian/witness/main/AGENTS.mdgit clone --depth 1 https://github.com/IngTian/witnessWrote 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/ingtian/witness/agents-md)<a href="https://agentmods.dev/instructions/ingtian/witness/agents-md"><img src="https://agentmods.dev/badge/instructions/ingtian/witness/agents-md/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/instructions/ingtian/witness/agents-md"><img src="https://agentmods.dev/badge/instructions/ingtian/witness/agents-md.svg" alt="Reviewed on agentmods" width="80" 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 | $0.02008 | $0.02008 |
| Opus 5 | $0.01004 | $0.01004 |
| Sonnet 5 | $0.00402 | $0.00402 |
| Haiku 4.5 | $0.00201 | $0.00201 |
Grade A, and why
witness 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 7d 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 — 125 lines — stays where its author put it; the contents beside it link to each section on GitHub.
AGENTS.md
Guidance for AI coding agents (and humans) working in this repo. Claude Code
reads this via @AGENTS.md in CLAUDE.md; other tools may read it
directly.
What this is
witness is a single pure-Go binary that captures Claude Code and OpenCode
sessions into a person-centric growth archive (raw turns → observations →
facets → narrative profile). It is a pure capture+serve tool — no coaching,
no injection into sessions. See README.md for the full model.
Build, test, run
Everything is CGO_ENABLED=0 (single static binary, no cgo). Use the Makefile:
make build # compile to bin/witness-<os>-<arch> (with version ldflags)
make test # CGO_ENABLED=0 go test ./...
make vet # CGO_ENABLED=0 go vet ./...
make fmt # gofmt -w internal cmd
make doctor # build + run the embedder/model/config health check
make install # (or: make install-opencode) wire hooks/plugin + MCP + bind runner
make build-all # cross-compile darwin+linux+windows, amd64+arm64
make package-windows # self-contained Windows zips (exe + prompts + model)
Before finishing a change, run make fmt vet test and expect all green.
macOS gotcha: always make build (or go build -o bin/...) — do not
cp a built arm64 binary to another path and run it. macOS invalidates the
ad-hoc code signature on copy and the copy dies with SIGKILL (137). Build in
place. (The Windows install path does copy the binary into %LOCALAPPDATA%
— that's fine, Windows has no such signature, and install_windows.go guards a
self-copy with os.SameFile + writes atomically via temp+rename.)
Architecture (where things live)
cmd/witness/— 15-line entry point; callscommands.Run().cmd/commands/— one cobra command per file (capture, session, worker, mcp, install, doctor, profile, facets, observations, review, lens, import, distill, cleanup). Shared helpers incli.go; terminal styling instyle.go.internal/store/— SQLite data model + the filesystem/DB-as-queue.raw(L0),observations(L1),facets/facet_versions(L2, bi-temporal),progress(distill watermark),staged,meta.MaxOpenConns(1)+ WAL +busy_timeoutis the deliberate concurrency model.resolveRoot()picks the data dir viadataDirNames(adopt-in-place, see invariants below).internal/distill/— the worker: mines L0→L1 per lens, reviews L1→L2, and regenerates L4 via a headless runner (claude -poropencode serve).internal/platform/{claude,opencode}/— the per-runtime capture/import adapters.internal/platform/opencode/plugin/embeds the OpenCode plugin JS. (internal/platform/platform.goholds the cross-runtime constants + types.)internal/bundle/— resolves bundled-asset dirs (prompts/, the embedding model) relative to the binary viaos.Executable()when no env var points at them, so a copied/installed binary self-locates its assets.internal/{embed,vector,lens,mcp}/— embeddings (pure-Go GoMLX), cosine search, lens loading, the MCP server.
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.
- 7d ago First seen · 125 lines · 2,008 tokens per session scan A b8c590c8ed45
witness AGENTS.md is an instructions file published in the GitHub repository IngTian/witness (8 stars, last pushed 15d ago), licensed MIT. It adds 2,008 tokens to every session, about $0.0100 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
next.js AGENTS.md
AGENTS.md instructions for vercel/next.js, covering next.js development guide, codebase structure, monorepo overview, core package: packages/next and other important packages.
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.
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).
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).
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.
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.