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/maaa2005/web-capability-compiler/agents-mdgit clone --depth 1 https://github.com/Maaa2005/web-capability-compilerWhat 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 | $0.00546 | $0.00546 |
| Opus 5 | $0.00273 | $0.00273 |
| Sonnet 5 | $0.00109 | $0.00109 |
| Haiku 4.5 | $0.00055 | $0.00055 |
Grade A, and why
web-capability-compiler 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 yesterday.
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 — 51 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Instructions for coding agents
The implementation source of truth is
docs/spec/WCC_AGENT_IMPLEMENTATION_SPEC_v0.2.1.md.
Read it before changing anything. This file summarizes Section 0 of that document;
where the two disagree, the specification wins.
Implement WCC in the phase order of specification Section 13. Do not broaden the scope without an explicit instruction from the project owner.
Rules
- Complete one phase and its acceptance criteria before starting the next phase.
- Prefer deterministic parsing over LLM inference whenever the same result can be obtained from HTML, ARIA, JSON-LD, or browser state.
- Do not implement irreversible actions such as purchase confirmation, account deletion, message sending, or state-changing form submission in MVP v0.1.
- Never execute a capability whose risk level is above
low. Higher levels may be listed but must be blocked. - Preserve evidence for every generated entity and capability. The system must be able to explain which DOM element or page signal produced each result.
- Add or update tests with every implementation change.
- Do not silently change schemas. Schema changes require a version increment and migration notes.
- Do not add an independent autonomous agent. WCC is a compilation and execution layer used by external agents.
- Keep the LLM provider optional and replaceable. Core extraction and execution must run without an LLM.
- All browser access goes through the
Driverinterface. Do not call Playwright APIs outsidesrc/wcc/driver/. - At the end of each phase, update
docs/progress.mdwith completed work, test results, unresolved issues, and the next phase. - Do not collapse
CapabilityTemplateandBoundCapabilityinto one model. Templates are portable and reusable; bound capabilities are session-scoped. - Do not treat a structural hash as a reusable page signature. Implement the three fingerprints of specification 7.3 with their specified roles.
- Do not add dynamic MCP tool registration in MVP v0.1. Use the six stable tools of specification Section 12.
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.
- yesterday First seen · 51 lines · 546 tokens per session scan A 2347ad14cac7
web-capability-compiler AGENTS.md is an instructions file published in the GitHub repository Maaa2005/web-capability-compiler (0 stars, last pushed 21d ago), licensed MIT. It adds 546 tokens to every session, about $0.0027 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
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.
deepseekCLI CLAUDE.md
Instructions for cynchro/deepseekCLI, covering tarea actual, progreso, decisiones clave, próximos pasos and contexto importante.
deckforge AGENTS.md
Instructions for tph-kds/deckforge, covering deckforge agent entry point, code intelligence, read order, default routing and non-negotiable implementation rules.
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).