flowstate
Skill2 starsChanged 6 months agoMITView on GitHub
Onboarding layer that maps the user's opening intent to relevant agent skills before work begins. Use when the user describes a new project, multi-step workflow, file-heavy setup, or asks which skills apply; also when re-evaluating mid-session after a major pivot. Reads references/skill-index.md and applies config/confidence-thresholds.md and config/verbosity.md.
2 other projects use a version of this.
Adopted by
Most recently taken up by
- dialectforge/FlowStateV1.1their own version, first seen Nov 1, 2026
- dialectforge/FlowStatetheir own version, first seen Oct 17, 2026
Other versions2 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 2
- dialectforge/FlowState★ 22same name
- dialectforge/FlowStateV1.1★ 22same 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 repositoryOne other file it gives its agents.
- .cursor/rules/flowstate.mdcCursor rule
The file
We split this skill into the rules it gives an agent
8 rules. About 0 are common, 6 are distinctive, and 2 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈6
Distinctive · estimate
This author's own choices. Worth a look.
2
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 44
If two skills compete, prefer the one with clearer task overlap and fewer anti-rule conflicts.
- Distinctive: line 47
If skill B typically follows skill A for this workflow, list A before B and optionally note *then* B in one line.
- Distinctive: line 52
Every listed skill gets a one-line reason tied to the user's wording.
- Very specific: line 56
After emitting the list, check which recommended skills do not already exist locally as SKILL.md files.
- Distinctive: line 59
Ask for approval once across all missing skills (not one prompt per skill).
- Very specific: line 62
If no match is found for a skill after searching, note it honestly: *"Couldn't find a public SKILL.md for <skill-name> — you may need to …
- Distinctive: line 66
If the user changes goal, stack, or deliverable in a way that invalidates earlier assumptions:
- Distinctive: line 70
Prefer a short delta (what changed in recommendations) unless verbosity is *compact* and the pivot is small.
≈ “Common” is an estimate: at least 1 in 5 of the 364,980 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
FlowState
Orchestrate skill discovery: interpret intent, score candidates, respect chains and anti-rules, then output a ranked, explainable list — and offer to fetch and install any missing skills with user approval.
When This Skill Applies
Activate when any of these are true:
- The user opens with project framing (I'm building…, I want to create…, Help me set up…).
- They describe a multi-step workflow across tools, file types, or domains.
- They upload several files at session start, signaling a non-trivial scope.
- They explicitly ask what skills or what to enable for a task.
- The conversation pivots materially (new goal, stack, or deliverable); re-run the mapping lightly (see Mid-session pivot).
Stay quiet for single-step, fully scoped requests where skill routing adds no value.
Operating Procedure
- Load config (do not show raw config to the user unless they ask):
config/verbosity.md— set response shape.config/confidence-thresholds.md— label each suggestion high / medium / maybe.
- Extract signals from the latest user message (and short thread context if needed):
- Nouns: languages, frameworks, file types, hosts (CI, cloud), artifacts (PDF, Excel).
- Verbs: migrate, scaffold, review, deploy, debug, split PRs, etc.
- Constraints: deadline, repo layout, team rules, readonly / no network.
- Match against
references/skill-index.md:- Primary: direct keyword or workflow overlap.
- Secondary: adjacent skills that often follow (see
skill-chains.md).
- Filter with
references/anti-skill-rules.md:- Drop skills that are explicitly ruled out for this context.
- If two skills compete, prefer the one with clearer task overlap and fewer anti-rule conflicts.
- Order using
references/skill-chains.md:- If skill B typically follows skill A for this workflow, list A before B and optionally note then B in one line.
- Score and label each remaining skill using confidence bands from
confidence-thresholds.md. - Emit in the shape required by
verbosity.md:- Every listed skill gets a one-line reason tied to the user's wording.
- End with a non-blocking question or invitation (e.g. whether to start with the top pick), not a hard gate.
- Offer to fetch missing skills:
- After emitting the list, check which recommended skills do not already exist locally as SKILL.md files.
- For each missing skill, search the web using the query:
"<skill-name>" SKILL.md cursor agent skill - Present the best match per skill to the user: name, source URL, and a one-line description of what was found.
- Ask for approval once across all missing skills (not one prompt per skill).
- On approval: write each fetched SKILL.md to
/skills/<skill-name>/SKILL.md. Do not overwrite skills that already exist unless the user explicitly asks. - On rejection: continue without installing; the user can ask later.
- If no match is found for a skill after searching, note it honestly: "Couldn't find a public SKILL.md for
<skill-name>— you may need to build it."
Mid-session Pivot
If the user changes goal, stack, or deliverable in a way that invalidates earlier assumptions:
- Re-scan signals from the pivot message; do not assume earlier skills still apply.
- Demote or remove skills that anti-skill-rules or the new goal rule out.
- Prefer a short delta (what changed in recommendations) unless verbosity is compact and the pivot is small.
- Re-run Step 8 for any newly recommended skills that are not yet installed.
Quality Bar
| Principle | Enforcement |
| Low friction | Cap list length per verbosity mode; no walls of text. |
| Ranked | Highest confidence first; optional maybe row only if useful. |
| Explainable | Each line ties to a user signal, not generic praise. |
| Non-blocking | User can ignore suggestions and continue. |
| Forward-looking | Include one secondary only when it clearly saves a later step. |
| Honest gaps | If a skill can't be found or installed, say so clearly. |
File Map
| Path | Role |
config/verbosity.md | Output length and structure. |
config/confidence-thresholds.md | Score → label mapping. |
references/skill-index.md | Canonical skill catalog and keywords. |
references/skill-chains.md | Sequencing and bundles. |
references/anti-skill-rules.md | Exclusions and de-prioritization. |
templates/quickstart-prompt.md | Optional paste-in for users who want a standard opener. |
evals/trigger-cases.md / pivot-cases.md | Regression-style scenarios for humans or automation. |
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.

