CourseForge
AGENTS.md0 starsChanged 55 days agoMITView on GitHub
CourseForge is a TypeScript/Node course-engineering harness: a deterministic 11-stage pipeline (concept → research → design → storyboard → editorial → visual direction → model → build → QA → release) that uses Claude Code or Codex for semantic work and code for routing, validation, rendering, QA and release decisions. Keep this file consistent with CLAUDE.md. src/cli + src/environment (CLI, doctor) · src/pipeline (state machine, stage loop, write audit, API) · src/routing (pure router) · src/harness (only place that knows the…
Where files are headingNew AGENTS.md files have more rules about keeping secrets out of the code, and fewer about code style.
Of the AGENTS.md files we read, the 31,774 first seen in the last six months, against the older ones: the kind of rule that gained most share, and the kind that lost most. About files like this in general, not this one, and an estimate, since a kind is a pattern in the words.
In this repository19 other files it gives its agents.
- CLAUDE.mdCLAUDE.md
- accessibility-reviewSkill
- alignment-reviewSkill
- assessment-designSkill
- assessment-reviewSkill
- course-gap-analysisSkill
- editorial-humanizationSkill
- fact-checkSkill
- html-intake-reviewSkill
- instructional-designSkill
- instructional-graphicsSkill
- research-authoringSkill
- research-planningSkill
- research-reviewSkill
- scenario-designSkill
- storyboard-authoringSkill
- ui-reviewSkill
- ux-reviewSkill
- visual-directionSkill
The file
We split this AGENTS.md into the rules it gives an agent
5 rules. About 1 is common, 1 is distinctive, and 3 are very specific to this project.
≈1
Common · estimate
Many files like this have a rule of this kind.
≈1
Distinctive · estimate
This author's own choices. Worth a look.
3
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 6
Keep this file consistent with CLAUDE.md.
- Very specific: line 34
Only src/core/proc.ts imports child_process (cross-spawn, shell: false, prompts on stdin).
- Very specific: line 38
Never edit input/originals/, versions/, locked artifacts or locked IDs.
- Very specific: line 39
Never read or modify ~/.claude, ~/.codex or global npm config.
- Common: line 40
No secrets, tokens, personal paths or emails anywhere in the repository.
≈ “Common” is an estimate: at least 1 in 5 of the 40,246 AGENTS.md files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
CourseForge: instructions for coding agents (Codex and others)
CourseForge is a TypeScript/Node course-engineering harness: a deterministic 11-stage pipeline (concept →
research → design → storyboard → editorial → visual direction → model → build → QA → release) that uses Claude
Code or Codex for semantic work and code for routing, validation, rendering, QA and release decisions.
Keep this file consistent with CLAUDE.md.
Sources of truth
docs/exec-plans/active/courseforge-v1.md: the approved plandocs/architecture/: the system as built;docs/adr/: recorded deviationsdocs/bootstrap/: original requirements of record (read-only)config/*.json: registries;src/core/schemas/: zod schemas for every artifact and agent output.claude/rules/architecture.md,courses.md,testing.md: rules that apply to you too
Map
src/cli + src/environment (CLI, doctor) · src/pipeline (state machine, stage loop, write audit, API) ·
src/routing (pure router) · src/harness (only place that knows the CLIs) · src/review + src/release ·
src/ingestion + src/artifacts · src/graphics + src/renderer + components/course-ui · src/qa ·
courses/<id>/ (one self-contained course each).
Commands
npm run build | npm run typecheck | npm run lint | npm test | npm run test:e2e | npm run test:acceptance
npm run gen:schemas # after changing src/core/schemas
node bin/courseforge.mjs <command> # CLI; see src/cli/help.tsNon-negotiables
- Provider-neutral core: only
src/harness/**names theclaude/codexCLIs, flags or output formats. - Only
src/core/proc.tsimportschild_process(cross-spawn,shell: false, prompts on stdin). - Deterministic routing from
config/viasrc/routing; agents return closed enums; plans saved first. - zod schemas are the source of truth; run
npm run gen:schemasand commitschemas/after changes. - Course isolation under
courses/<id>/usingsrc/core/fsx.tshelpers. - Never edit
input/originals/,versions/, locked artifacts or locked IDs; never bypass the write audit. - Never read or modify
~/.claude,~/.codexor global npm config; no global installs. - No secrets, tokens, personal paths or emails anywhere in the repository.
Parallel work
Stay inside your assigned directories. Only the coordinator edits package.json, the lockfile, tsconfig*.json,
biome.json, runner configs and shared interfaces (src/core/*, src/harness/types.ts,
src/pipeline/api.ts). Do not install, commit, stash or check out unless asked; do not run e2e alongside other
browser or agent work.
Skills (index)
Codex has no native project-skills path. These files are the single source used by CourseForge's prompt assembly and by Claude Code; read the relevant one when doing that kind of work:
| Skill | File |
| research-planning | .claude/skills/research-planning/SKILL.md |
| research-authoring | .claude/skills/research-authoring/SKILL.md |
| research-review | .claude/skills/research-review/SKILL.md |
| fact-check | .claude/skills/fact-check/SKILL.md |
| instructional-design | .claude/skills/instructional-design/SKILL.md |
| alignment-review | .claude/skills/alignment-review/SKILL.md |
| storyboard-authoring | .claude/skills/storyboard-authoring/SKILL.md |
| scenario-design | .claude/skills/scenario-design/SKILL.md |
| assessment-design / assessment-review | .claude/skills/assessment-design/SKILL.md, .claude/skills/assessment-review/SKILL.md |
| editorial-humanization | .claude/skills/editorial-humanization/SKILL.md |
| course-gap-analysis | .claude/skills/course-gap-analysis/SKILL.md |
| visual-direction / instructional-graphics | .claude/skills/visual-direction/SKILL.md, .claude/skills/instructional-graphics/SKILL.md |
| html-intake-review | .claude/skills/html-intake-review/SKILL.md |
| accessibility-review / ui-review / ux-review | .claude/skills/accessibility-review/SKILL.md, .claude/skills/ui-review/SKILL.md, .claude/skills/ux-review/SKILL.md |
Prompt templates: prompts/templates/*.md ({{variable}} substitution only). Reviewer rubrics:
prompts/rubrics/<reviewer-id>.md. Reviewer and skill routing: config/reviewers.json, config/skills.json.
Before you say you are done
npm run typecheck, npm run lint and npm test pass; npm run test:e2e too if you touched the renderer,
components, runtime, graphics or QA; npx tsx scripts/gen-schemas.ts --check after schema changes; tests that
cover a requirement carry its acceptance tag (@C6 …); docs/ still matches the code.
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.

