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 skills/ucsahinn/codex-chef/fetchnpx skills add ucsahinn/codex-chef --skill fetchgit clone --depth 1 https://github.com/ucsahinn/codex-chefWhat 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 | $0.00075 | $0.01278 |
| Opus 5 | $0.00037 | $0.00639 |
| Sonnet 5 | $0.00015 | $0.00256 |
| Haiku 4.5 | $0.00007 | $0.00128 |
Grade A, and why
fetch 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 3d 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 — 119 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Fetch
Reconstruct the observable website experience in the current workspace. A URL is a reference target, not permission to extract private data, server source, credentials, sessions, or protected assets.
Invocation
- Use
$fetch <url>when Codex Chef installed the managed direct skill. - Use
$codex-chef-workflows:fetch <url>when invoking the installed plugin package directly. - Keep implicit invocation disabled on both surfaces. A normal request that happens to mention fetching, scraping, or the JavaScript Fetch API must not activate this workflow.
Input Contract
- Require a valid
https://reference URL. Allowhttp://, localhost, an intranet host, or a non-standard port only when the user explicitly names that exact target and the active network policy permits it. - Use the current workspace as the output target when it contains one clear application. If several apps or no safe output location exists, ask only for the target path.
- Treat a bare public URL as authority only for bounded, read-only inspection of public, unauthenticated pages and a local reconstruction with inert auth and lawful assets. Exact protected branding, private content, or publication still requires the user to confirm ownership or permission.
- Require explicit authorization before inspecting authenticated or private routes, using a test account, calling a non-public API, or writing outside the current workspace.
- Read safety-boundaries.md before any browser or network inspection.
Workflow
- Route the work
- Inspect repo instructions, scripts, stack, and dirty state first.
- Keep implementation in the main thread. Use
code_mapper,design_reviewer,frontend_verifier, orsecurity_auditoronly for independent mapping, visual review, browser verification, or auth risk. - Prefer a real browser through Playwright or Chrome DevTools. Raw HTTP downloads are supplemental because they miss client-rendered states.
- Capture the reference
- Follow capture-protocol.md.
- Establish passive request and destination guards before the first navigation. If the selected browser tool cannot do that, stop as blocked.
- Inventory routes, viewports, states, navigation, forms, assets, fonts, motion, console behavior, and sanitized request shapes.
- Capture source screenshots before implementation. Do not use remembered appearance as evidence.
- Build the reconstruction
- Follow implementation-protocol.md.
- Preserve the existing stack and conventions when a project already exists. Do not install or upgrade dependencies without approval.
- Implement working local behavior, responsive states, loading, empty, error, focus, hover, and disabled states that were observed.
- Replace unavailable server behavior with a clearly local fixture or mock. Never silently proxy a clone to a production mutation endpoint.
- Verify in a real browser
- Follow verification-rubric.md.
- Compare source and local screenshots at matching viewport, route, scroll position, and UI state. Exercise navigation and interactions.
- Fix material mismatches and rerun the narrowest relevant checks.
- Prove completion
- Create a fidelity report from fetch-report.template.json.
- Run:
What ships with it
9 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
- agents/openai.yaml 394 B
- assets/fetch-report.template.json 4.3 KB
- references/capture-protocol.md 5.2 KB
- references/forward-tests.md 3.3 KB
- references/implementation-protocol.md 3.7 KB
- references/safety-boundaries.md 6.4 KB
- references/sources.md 2.8 KB
- references/verification-rubric.md 3.2 KB
- scripts/validate-fetch-report.mjs 32 KB runs code
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.
- 3d ago First seen · 119 lines · 75 tokens per session scan A aa1aa7b13e64
fetch is a skill published in the GitHub repository ucsahinn/codex-chef (6 stars, last pushed 16d ago), licensed MIT. It adds 75 tokens to every session and 1,278 once invoked, about $0.0004 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 skills, from other repositories
mermaid-visualizer
Transform text content into professional Mermaid diagrams for presentations and documentation. Use when users ask to visualize concepts, create flowcharts, or make diagrams from text. Supports process flows, system architectures, comparisons, mindmaps, and more with built-in syntax error prevention.
apm-usage
Activate when the user asks about APM (Agent Package Manager): installing, configuring, authoring, or troubleshooting AI-agent packages, dependencies, compilation, MCP servers, policy, or any apm CLI command.
clean-user-facing-text
Audit and finalize authorized natural-language text meant for readers: strip suspicious invisible Unicode, then rewrite prose while keeping facts, meaning, and the writer's voice. Use when the user asks to clean, humanize, polish, or finalize articles, manuscripts, reports, documentation, emails, product copy, UI…
sessions
Search and ask questions about coding agent session history across Claude Code, Codex, and Cursor. Use when asking what was worked on, what was tried before, how a problem was investigated across sessions, what happened recently, or any question about past agent sessions. Also use when the user references prior…
codex-autoresearch
Run autonomous, measurable experiments in a Git repository: change one hypothesis, verify a numeric metric, keep improvements, and revert failures. Use when the user wants Codex to keep iterating toward a numeric target in the foreground or as a detached background run. Do not use for ordinary one-shot coding…
map-review
Interactive 4-section code review using monitor, predictor, and evaluator agents plus the user and maintainer role reviewers on current changes. Use when reviewing a diff, PR, or staged work before merge. Do NOT use to plan or implement; use map-plan or map-efficient.