WebCodex is software that lets AI coding agents work with code and developer tools on a user's own machine. It is used to inspect and edit repositories, run commands and tests, use Git, and investigate failures without moving the project to a hosted workspace. The catalogue entry provides an instruction for using WebCodex.
Borrowing it
Nothing to install: this file belongs to yyjeqhc/webcodex. 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/yyjeqhc/webcodex/main/AGENTS.mdgit clone --depth 1 https://github.com/yyjeqhc/webcodexWrote 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/yyjeqhc/webcodex/agents-md)<a href="https://agentmods.dev/instructions/yyjeqhc/webcodex/agents-md"><img src="https://agentmods.dev/badge/instructions/yyjeqhc/webcodex/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/yyjeqhc/webcodex/agents-md"><img src="https://agentmods.dev/badge/instructions/yyjeqhc/webcodex/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 | $0.01928 | $0.01928 |
| Opus 5.5 | $0.00771 | $0.00771 |
| Sonnet 5.5 | $0.00386 | $0.00386 |
| Haiku 4.5 | $0.00193 | $0.00193 |
Grade A, and why
webcodex 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 15d 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 — 70 lines — stays where its author put it; the contents beside it link to each section on GitHub.
AGENTS.md — WebCodex Repository Rules
Use the host's tools for local WebCodex work and CONTRIBUTING.md for setup/validation. Deeper AGENTS.md files govern their directories. Load domain guidance at the relevant boundary; webcodex.workflow applies to connected-runtime callers.
WebCodex is actively developed. Features, fixes, and reliability work are welcome, but preserve credential, process-tree, transport, durability, authority, replay-safety, and boundedness contracts.
1. Product and authority invariants
- Protect real authority boundaries: credentials, public entry points, destructive actions, execution targets, process ownership, history, and published artifacts. Model new credential audiences, scopes, identities, leases, and replaceable targets explicitly; stale requests must not retarget or gain authority.
- Never print, commit, or expose credentials, authorization headers, private keys, tokens, secret files, or sensitive command output.
- Structured execution lifecycle state owns retry safety. Semantic ambiguity, authority changes, stale fences, replay uncertainty, and uncertain effects fail closed; presentation or model ergonomics never override execution authority.
- Keep one canonical machine representation per fact or follow-up. Static/internal taxonomies and forensic bookkeeping belong in runtime types, tests, telemetry, or exceptional diagnostics rather than duplicate model-facing fields. Model-facing compatibility requires a named concrete consumer; historical aliases are not retained merely because an older implementation emitted them.
- HTTP/MCP requests are stateless unless the exact adapter contract supplies a stable
ClientWindow. Never infer Workflow Session, model-context retention, or authority from transport/audit IDs, credentials, Projects, or prior requests. - Keep Connector Tasks, Workflow Sessions, durable Agents/Conversations/Wakes, and Agent Tasks distinct. Persistence of identity or history is not permission to replay an effect.
- Keep core execution and Job semantics protocol-, UI-, transport-, and OS-neutral. Host-specific behavior, including MCP App orchestration, belongs in explicit adapters.
- Boundedness is a correctness property for model surfaces, logs, persistence, scans, queues, and retained state. Do not turn ergonomic limits into authority bypasses or silently widen resource contracts.
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.
- 15d ago Changed · +3 lines · +112 tokens per session 45223b0151e0
- 26d ago Changed · +2 lines · +120 tokens per session 4cb9112a059e
- 1mo ago Changed · +1 lines · +60 tokens per session 91bc5cd88add
- 1mo ago Changed · -15 lines · -750 tokens per session dccfd6b26342
- 1mo ago Changed · -2 tokens per session 80f063e8411c
- 1mo ago Changed · +48 tokens per session b61a02d7d7d2
- 1mo ago Changed cfdaa1b9ac4d
- 1mo ago Changed · +66 tokens per session b7318d178c99
- 1mo ago Changed · +1 lines · +61 tokens per session 11008c3bc51f
- 1mo ago First seen · 78 lines · 2,213 tokens per session scan A dcda081d3cca
webcodex AGENTS.md is an instructions file published in the GitHub repository yyjeqhc/webcodex (2,488 stars, last pushed 2d ago), licensed Apache-2.0. It adds 1,928 tokens to every session, about $0.0077 per session on Opus 5.5. 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.
webcodex AGENTS.md compared with instructions from other repositories
| Mod | Type | Stars | Last push | Licence | Scan | Tokens | Works with |
|---|---|---|---|---|---|---|---|
| webcodex AGENTS.mdyyjeqhc/webcodex | Instructions file | 2.5k | 2d ago | Apache-2.0 | A | 1,928 | Codex, OpenCode |
| vscode buildNext.instructions.mdmicrosoft/vscode | Instructions file | 195k | yesterday | MIT | A | 9,719 | GitHub Copilot |
| next.js AGENTS.mdvercel/next.js | Instructions file | 143k | today | MIT | A | 7,461 | Codex, OpenCode |
| langchain AGENTS.mdlangchain-ai/langchain | Instructions file | 149k | yesterday | MIT | A | 4,680 | Codex, OpenCode |
| vscode oss-third-party-notices.instructions.mdmicrosoft/vscode | Instructions file | 195k | yesterday | MIT | A | 5,516 | GitHub Copilot |
| vscode copilot-instructions.mdmicrosoft/vscode | Instructions file | 195k | yesterday | MIT | A | 3,093 | GitHub Copilot |
| deepseek-harness AGENTS.mddeepseek-ai/deepseek-harness | Instructions file | 274k | 2d ago | MIT | A | 4,313 | Codex, OpenCode |