context-engineering-protocol
AGENTS.md9 starsChanged 3 months agoApache-2.0View on GitHub
Instructions for AI coding agents (e.g. OpenAI Codex) working in this repository. This repository packages its capabilities as portable "skills" - read the relevant skill's SKILL.md in full and follow it before acting in the corresponding area. Do not guess at a skill's behavior from this table alone.
In this repository22 other files it gives its agents.
- .cursor/rules/compiling-project-guidelines.mdcCursor rule
- .cursor/rules/demo-consume-context.mdcCursor rule
- .cursor/rules/demo-write-user-stories.mdcCursor rule
- .cursor/rules/ult-autoscaffold-content.mdcCursor rule
- .cursor/rules/ult-cep-retrofit.mdcCursor rule
- .cursor/rules/ult-cep-wizard.mdcCursor rule
- .cursor/rules/ult-codegraph.mdcCursor rule
- .cursor/rules/ult-context-generate.mdcCursor rule
- .cursor/rules/ult-institutional-memory-distill.mdcCursor rule
- .cursor/rules/ult-onboarding-index.mdcCursor rule
- .cursor/rules/ult-repo-layout.mdcCursor rule
- compiling-project-guidelinesSkill
- demo-consume-contextSkill
- demo-write-user-storiesSkill
- autoscaffold-contentSkill
- cep-retrofitSkill
- cep-wizardSkill
- codegraphSkill
- context-generateSkill
- institutional-memory-distillSkill
- onboarding-indexSkill
- repo-layoutSkill
The file
Files
AGENTS.md
AGENTS.md
Instructions for AI coding agents (e.g. OpenAI Codex) working in this repository.
This repository packages its capabilities as portable "skills" - read the
relevant skill's SKILL.md in full and follow it before acting in the
corresponding area. Do not guess at a skill's behavior from this table alone.
| Skill | Description | Path |
| compiling-project-guidelines | Compile scattered guideline sources into one scope-aware COMPILED-GUIDELINES.md for other skills and ult-context-generate's Constraints layer. Do NOT use to enforce rules at runtime. | .github/skills/compiling-project-guidelines/SKILL.md |
| demo-consume-context | Worked example that discovers, loads, and tags a context package per CONSUMING-CONTEXT-PACKAGE.md, then writes a reverse-index addendum — proves the produce/consume/tag loop end-to-end. Do NOT use for real feature work. | .github/skills/demo-consume-context/SKILL.md |
| demo-write-user-stories | Reference implementation that writes functional user stories from a feature description, optionally grounded in an approved CEP context package per CONSUMING-CONTEXT-PACKAGE.md, with per-story citations and the standard tag/reverse-index loop. A worked example for downstream teams building their own context-consuming skills. | .github/skills/demo-write-user-stories/SKILL.md |
| ult-autoscaffold-content | Generate real starter content for a project's What-L2 (requirements) and How-L2 (architecture/conventions) CEP layers once ult-repo-layout has resolved their paths but found them empty — an honest, minimal, YAML-frontmatter-first overview document per layer for small/single targets, or graphify-informed per-module tiering with resumable per-module CONTEXT.md generation, existence-gated CODING-STANDARDS.md/TESTING-GUIDELINES.md, interface-boundary docs for graph-crossing module pairs, and a rendered CEP-INDEX.md router for large repos, optionally informed by a user-supplied domain-pack of terminology/references if one is configured. This skill generates a NEW starter doc only when none exists — it never compiles or reconciles EXISTING scattered guideline sources (that's compiling-project-guidelines' job). Do NOT use to enforce layout paths or run layer discovery — that's ult-repo-layout. Do NOT use to author or generate a domain pack — this skill only ever consumes one you already wrote. | .github/skills/ult-autoscaffold-content/SKILL.md |
| ult-cep-retrofit | Point at an existing skill library, inventory it format-agnostically, and (with human confirmation at every step) insert pointers to CEP's existing CONSUMING-.md contracts into the skills that should consume them. Do NOT use to write new CONSUMING-.md contracts, validate that a retrofit works at runtime, or edit a library's own logic beyond the inserted pointer. | .github/skills/ult-cep-retrofit/SKILL.md |
| ult-cep-wizard | Launch a local, localhost-only browser wizard with two journeys — layout onboarding for a project with ult-repo-layout installed, initialized or not — guides an uninitialized or not-yet-discovered repo through Run Discover, then shows resolved layer/slot state through four labeled boxes and a directory picker, and lets you resolve pending layout decisions (confirm/skip/disable/pick-a-directory) and Apply them into context-config.yaml via ult-repo-layout's own confirm step; and retrofit — walks an existing skill library through ult-cep-retrofit's inventory/classify/draft/apply flow, with an editable per-file diff preview and batch Apply, never writing without that preview. Do NOT use for headless/CI-only layout validation — use ult-repo-layout's discover/confirm-layers/--validate directly. | .github/skills/ult-cep-wizard/SKILL.md |
| ult-codegraph | Generate a codebase knowledge graph with graphify at graphify-out/ so other skills can query cross-file relationships before touching code. Do NOT use for runtime profiling. | .github/skills/ult-codegraph/SKILL.md |
| ult-context-generate | Assemble a context package (code graph, requirements, constraints, blast radius) before a downstream generation task runs - human-approved, source-attributed. Do NOT use for simple lookups. | .github/skills/ult-context-generate/SKILL.md |
| ult-institutional-memory-distill | Distill decisions, reasoning, and rejected alternatives from PRs, design docs, and postmortems into the project's decision_ledger, so ult-context-generate's trip-wire can surface institutional memory before new work quietly repeats settled ground. Do NOT use to query the ledger against new work or decide revise/proceed/escalate -- that's ult-context-generate/SKILL.md Step 7.7's job. | .github/skills/ult-institutional-memory-distill/SKILL.md |
| ult-onboarding-index | Discover CEP-managed content already present in a target repo (What/How layer docs, compiled guidelines, context packages, decision ledger) via existence checks against layout-slots-registry.yaml-resolved paths, then write one canonical root AGENTS.md onboarding index plus thin per-tool pointer stubs (.github/copilot-instructions.md, CLAUDE.md, .cursor/rules/onboarding.mdc) that link into it. All writes are marked-block or existence-gated and non-destructive, including install.sh's own AGENTS.md skill-catalog block. Codex needs no separate stub -- root AGENTS.md is already its native onboarding format. Do NOT use it to generate the CEP content itself (ult-autoscaffold-content/compiling-project-guidelines do that) or to resolve layer paths (ult-repo-layout does that). Do NOT confuse it with this OSS repo's own AGENTS.md (generated by catalog/export_adapters.py) -- this skill only ever writes into a separate target repo being onboarded with CEP. | .github/skills/ult-onboarding-index/SKILL.md |
| ult-repo-layout | Register, resolve, and validate where a project's path-slots actually live via .layout-slots.yaml markers, so relocating a slot needs zero SKILL.md edits. Do NOT use for single-file lookups. | .github/skills/ult-repo-layout/SKILL.md |
Generated by catalog/export_adapters.py from each skill's SKILL.md frontmatter - do not hand-edit; run python catalog/export_adapters.py --write after adding or changing a skill.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.