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/agentpilled/clipbrain/agents-mdgit clone --depth 1 https://github.com/agentpilled/ClipBrainWhat 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 | $0.01011 | $0.01011 |
| Opus 5 | $0.00505 | $0.00505 |
| Sonnet 5 | $0.00202 | $0.00202 |
| Haiku 4.5 | $0.00101 | $0.00101 |
Grade A, and why
ClipBrain 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 2d 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.
curl -sS http://127.0.0.1:19285/health How it starts
The opening of the file, as written. The whole thing — 89 lines — stays where its author put it; the contents beside it link to each section on GitHub.
ClipBrain Agent Instructions
This file is for Codex and other coding agents working in this repo. Claude Code
can use CLAUDE.md; keep this file shorter and focused on operational rules.
Project Shape
ClipBrain is a Chrome extension plus a local Bun HTTP server. The extension
captures web pages, Kindle highlights, Gmail threads, YouTube transcripts, and
PDFs. The server writes captures to the user's local knowledge base through the
installed gbrain CLI and optionally syncs markdown to Obsidian.
Important files:
server.ts: local HTTP API, dashboard routes, capture handling, diagnosticsclipbrain-mcp.ts: ClipBrain-specific MCP tools that call the local serverbackfill.ts: controlled Knowledge Compiler backfill for legacy capturescorpus-report.ts: read-only corpus quality report for junk/duplicate auditcleanup-plan.ts: read-only cleanup recommendations; never deletes or writescleanup-apply.ts: explicit approval-gated cleanup executor with backups/logsgbrain-list.ts: shared robustgbrain listscanner used by server, corpus, and cleanup flowsdoctor.ts: first-run diagnostics for launch readiness and local setup issuestwitter-agent.ts: draft-only building-in-public generator; never posts to Xservice-worker.js: MV3 background worker, capture dispatch, offline queuecontent-script.js,kindle-content-script.js,gmail-content-script.js: page extractorspost-process.ts: optional OpenAI enrichment after a capture is saveddashboard.html,popup.html,popup.js: local UI surfacessetup.sh,setup-mcp.sh: local install and MCP configurationtest/: Bun tests for server and post-processing behavior
Runtime Facts
- Use Bun. Primary checks are
bun run doctor,bun test,bun audit, and shell syntax checks for setup scripts. - The server binds to
127.0.0.1:19285by default. It can be overridden with--host,--port,GBRAIN_CAPTURE_HOST, andGBRAIN_CAPTURE_PORT. - MCP setup registers both
gbrain(gbrain serve) andclipbrain(bun clipbrain-mcp.ts). Keepgbrainfor broad brain tools; useclipbrainfor ClipBrain-specific agent handoffs such ascontext_pack. - Do not reintroduce a vendored
gbraindependency or local./bin/gbrainbuild path. The app should useGBRAIN_BINwhen set, thengbrainfromPATH. package-lock.jsonis intentionally absent. This repo usesbun.lock.- Runtime data files such as
.captures.jsonl,.highlight-count,.highlight-counts.json,.clipbrain.json,.cleanup-actions.jsonl, and.cleanup-backups/must stay out of commits. Tests should useCLIPBRAIN_DATA_DIRfor isolated state. - Use
bun run backfill --limit Nfor a dry-run before applying compiler upgrades withbun run backfill --apply --limit N. - Use
bun run corpusto inspect likely junk, duplicate titles, and pending compiler upgrades. It is read-only. - Use
bun run cleanup-planto turn the corpus audit into a read-only keep/delete/merge/title-fix proposal before any manual cleanup. - Use
bun run cleanup-applyfor a dry-run of executable cleanup operations. Never run it with--executeunless the user has approved exact--approve action:slugtokens.
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.
- 2d ago First seen · 89 lines · 1,011 tokens per session scan A f5a2a3e0c350
ClipBrain AGENTS.md is an instructions file published in the GitHub repository agentpilled/ClipBrain (24 stars, last pushed 2mo ago), licensed MIT. It adds 1,011 tokens to every session, about $0.0051 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
Browser4 CLAUDE.md
Instructions for platonai/Browser4, covering browser4 — project context for claude, architecture, key dispatch chain (cli → browser), batch commands and e2e test structure.
fast-browser CLAUDE.md
Claude Code instructions for m4ttstack/fast-browser, covering fast browser plugin, where a change belongs, fork branch: use fast-browser-runtime, releasing a new runtime and re-pinning this repo: use the script.
agentoscope AGENTS.md
Instructions for rafaelcg/agentoscope, covering agentoscope — agent instructions, what this is, layout, commands and conventions.
sendblue-browser-use AGENTS.md
Instructions for sendblue-api/sendblue-browser-use, covering agents.md — sendblue-browser-use, what this repo is, setup, common commands and health (no auth).
statbotics-mcp-server CLAUDE.md
Claude Code instructions for withinfocus/statbotics-mcp-server, covering claude instructions for statbotics mcp server, project overview, project structure, source code architecture and test architecture.
sniff AGENTS.md
Instructions for Aboudjem/sniff, covering agents.md: sniff, what this repo is, how an agent should use sniff, handling the playwright setup gate and finding output schema.