Generated from .claude/skills/local/todo-manager.md — do not edit here.

✎ Edit this page’s source

This is a stub, not the skill. The skill is Session Task Manager (folio-core), from cat-harness/skills/sdlc/sdlc-core, which is what skill_fetch serves. Read that one; this page exists only so an old link still lands somewhere truthful.

todo-manager — see skills/sdlc/sdlc-core/

This is a stub. The skill lives in the kg graph, not here.

skill_fetch("todo-manager") serves skills/sdlc/sdlc-core/todo-manager.md — LOCAL_PACKAGES in src/tools/skill-fetch.ts holds skills/sdlc/sdlc-core and no .claude/skills/local entry — so that file is what an agent asking for this skill by name has always received.

A 369-line hand-authored copy sat here until 2026-09-19 and diverged from the servable one by 261 diff lines. It was not servable, carried no front matter, and lacked ## Check before you create — the STRICT rule that exists because an unguarded beans create produced 14,688 duplicate beans. Meanwhile the onboarding guide, in five languages, pointed here as the “full discipline”. Four rules it carried and the servable copy did not were ported across first; nothing was dropped. Bean tdmg.

The skill is now three, because the one file had grown to 396 lines against a 400-line ceiling while carrying three separable disciplines:

skill	what it governs
todo-manager	bean mechanics: the CLI, the STRICT check before beans create, the fallback when the CLI is absent, the status vocabulary, coordination
opening-brief	what you say before starting a bean or a topic
turn-reporting	what you say during and after each turn, including the STRICT “next”-line rule

Same discipline as CLAUDE.md and GEMINI.md: one source of truth, thin pointers to it. Do not re-add content here — edit the skill.

Processes that run this skill

process	step(s) that name it
CRDM Phase 5 — beans and sign-off	Phase 5: Create beans
Agent bean lifecycle	Check before you create (exact-title search); Create the bean (agent CLI, not an engine op); Work, keeping the body current (this is ‘edit’); Complete (no unchecked todos left); Scrap with reasons NEVER delete
Code change and review	Record what was done, and close
Is the incremental IG AST what a full build would have produced?	Note the missed coupling on the bean
Incremental IG build	Log the environment error on the bean; Log findings on the bean; File QC findings as beans
L3 FHIR IG pipeline	File QC findings as beans
Getting started	Seed the work plan
Authoring a document	2 · Seed the work plan
Content Change and Review	Open the branch-watch bean; Note the main-branch watch
Content lifecycle	Seed the work plan; File feedback as beans
Draft, review and publish	Open or claim the release bean; Open beans for the change requests; Close the release beans; Archive the release bean
Editing and HCI validation	Claim or open the bean; Log findings on the bean; Resolve or re-open the bean
Evidence for a recommendation	Open a bean for the unverified citation; Record the evidence gap
L1 document ingestion — a document to the L1 source knowledge graph	Record the gap as a bean
Authoring a paper	2 · Seed the work plan
L2 component edit and review	Reopen upstream, flag derived assets
L2 DAK authoring	Seed the work plan
Phase 1: Planning	Assign work and begin sprint cycle