flow-learn
Skill19 starsChanged 5 days agoApache-2.0View on GitHub
Use when reorienting across machines or harnesses (/flow:refresh), reconciling spec checklists (/flow:sync), auditing codebase or spec drift, sharpening domain glossaries and ADRs, or resynthesizing OKF v0.2 knowledge and .agents/skills/ (/flow:learn, /flow:resynthesize).
1 other project uses a version of this.
Adopted by
Most recently taken up by
- baodhg/FLOWtheir own version, first seen Nov 8, 2026
Other versions1 project uses a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 1
- baodhg/FLOW★ 2same name
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository10 other files it gives its agents.
- AGENTS.mdAGENTS.md
- debloatSkill
- flow-completionSkill
- flow-docsSkill
- flow-executionSkill
- flow-planningSkill
- flow-setupSkill
- flowSkill
- multi-harness-pluginsSkill
- okfSkill
The file
We split this skill into the rules it gives an agent
7 rules. About 0 are common, 0 are distinctive, and 7 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈0
Distinctive · estimate
This author's own choices. Worth a look.
7
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 13
Use for /flow:learn [scope] (and aliases resynthesize, refresh, sync, docs).
- Very specific: line 28
When resynthesizing >= 2 domains, fan out 1 isolated synthesis subagent per domain (type_name="self" on Google Antigravity) so each domain …
- Very specific: line 31
Follow ../multi-harness-plugins/SKILL.md to update .agents/skills/<domain>/SKILL.md (<= 200 lines + references/*.md), …
- Very specific: line 32
When updating user-facing guides, READMEs, or module docs, follow ../flow-docs/SKILL.md.
- Very specific: line 33
Remove any legacy .agents/skills/flow-memory-keeper/ or copied global Flow skills.
- Very specific: line 39
Never overwrite uncommitted user code changes or developer-owned content inside <!-- project-customization: start --> blocks.
- Very specific: line 41
Every knowledge claim must cite a verified repository path (path/to/file::Symbol) or primary source in sources:.
≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill 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
Flow Learn (Drift True-Up, Cross-Harness Refresh, & OKF Resynthesis)
<!-- lifecycle-ownership: owner=flow-learn; operations=learn,resynthesize,refresh,sync,docs -->Trigger
Use for /flow:learn [scope] (and aliases resynthesize, refresh, sync, docs). Can run at any point in a project's lifecycle—when switching machines or harnesses, mid-flow, post-flow, or standalone. Follow ../flow/references/learn.md.
Workflow
- Cross-Machine/Harness Reorientation & Multi-Chapter Drift Audit (
refresh+sync+true-up):- Spec, Chapter, & Task Reconciliation (
sync/refresh):- Compare
tasks/*.mdfrontmatter (state,commit,claimed_by),chapters/<ch_id>/spec.md(state,planning_status,last_reviewed_commit),chapters/index.md, and rootspecs/<flow_id>/spec.md(chapters:,current_chapter, checklist markers,## 6. Current State & Remaining Plan Snapshot) against livegit logand working-tree state. - Mark tasks whose commits landed on another machine/session
state: completedwith their verified GitcommitSHA, release stale cross-harnessin_progressclaims when no active work matches the working tree, and rewritechapters/index.md, chapterspec.md, and rootspec.mdchecklists and state snapshots in place in present tense. - Flag any open
tasks/<id>-<slug>.mdor downstream chapters whose targetfiles:orpath/to/file::Symbolsignatures drifted so/flow:revise <flow_id> [chapter_id]can re-baseline them.
- Compare
- Source-to-Knowledge Drift (
sources[].resourceIntersection):- Intersect modified or deleted files/symbols (
git status,git diff <last_verified_commit>..HEAD) against every.agents/bundles/knowledge/**/*.mdchapter'ssources[].resourcepaths andpath/to/file::Symbolanchors (sharding across1read-only subagent per domain on multi-domain repos;type_name="self"on Google Antigravity).
- Intersect modified or deleted files/symbols (
- Manifest & Toolchain Drift:
- Compare
pyproject.toml,Cargo.toml,package.json,go.mod,Makefile, and CI workflows againstproduct/tech-stack.md,product/product.md, andknowledge/workflow.md.
- Compare
- Spec, Chapter, & Task Reconciliation (
- Subagent-Sharded OKF v0.2 Knowledge Resynthesis (
.agents/bundles/knowledge/):- Harvest confirmed discoveries across single-chapter and nested
chapters/<ch_id>/tasks/*.md(## Notes & Discoveries) and promotedresearch/*.md. - When resynthesizing
>= 2domains, fan out 1 isolated synthesis subagent per domain (type_name="self"on Google Antigravity) so each domain chapter receives thorough present-tense updates withstatus: stable,<= 40-line<!-- truth: start -->blocks,sources:+[^id]:footnotes, and zero standalone---body rules. - Sharpen
knowledge/domains/<domain>.mdand apply the ADR Filter (hard to reverse + surprising without context + genuine trade-off) forknowledge/decisions/adr-NNN-<slug>.md, pruning chronological sediment and model-relative no-ops (../flow-docs/references/agent-context.md).
- Harvest confirmed discoveries across single-chapter and nested
- Dynamic Project-Local Skill, Subagent, & Documentation Synchronization (
.agents/skills/,.agents/agents/, &docs/):- Follow
../multi-harness-plugins/SKILL.mdto update.agents/skills/<domain>/SKILL.md(<= 200lines +references/*.md),.agents/skills/review-<domain>/SKILL.md(references/checklist.md+references/false-positive-suppressions.md),.agents/skills/index.md, paired subagents (.agents/agents/<domain>-reviewer.mdand<domain>-specialist.mdwithcommandExecutionPolicy: sandbox), and rootAGENTS.md. - When updating user-facing guides, READMEs, or module docs, follow
../flow-docs/SKILL.md. - Remove any legacy
.agents/skills/flow-memory-keeper/or copied global Flow skills.
- Follow
- Frontmatter-Free Index Rebuild,
log.mdAudit Entry & True-Up Dashboard:- Rebuild root
.agents/bundles/index.md(okf_version: "0.2") and frontmatter-free subdirectoryindex.mdfiles (product/index.md,knowledge/index.md,knowledge/<ns>/index.md,specs/index.md,chapters/index.md), prepend asynthesize/verifyentry to.agents/bundles/log.md, and report the Project True-Up & Reorientation Dashboard.
- Rebuild root
Guardrails
- Never overwrite uncommitted user code changes or developer-owned content inside
<!-- project-customization: start -->blocks. - Rewrite updated knowledge chapters in place in present tense; never append chronological revision logs inside concept documents.
- Every knowledge claim must cite a verified repository path (
path/to/file::Symbol) or primary source insources:.
Output
Return the Project True-Up & Reorientation Dashboard listing reconciled commits/claims, repaired knowledge/** chapters, updated .agents/skills/ and .agents/agents/, and the next exact step.
Validation
Confirm all sources[].id entries have matching [^id] footnotes in body prose, all relative links in .agents/bundles/index.md and .agents/skills/index.md resolve on disk, spec.md checklists match tasks/*.md frontmatter, and .agents/bundles/log.md records the synthesis entry.
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.

