flow
Skill1 starsChanged 5 days agoApache-2.0View on GitHub
Run the full recon → spec → implement → review pipeline for one task, each step in its own subagent to keep the main conversation's context small, pausing when a step needs input and stopping cleanly if one fails.
34 other projects use a version of this.
Adopted by
Most recently taken up by
- manhquydev/flow-skilltheir own version, first seen Nov 15, 2026
- retospect/precis-mcptheir own version, first seen Nov 15, 2026
- marco-bertelli/filotheir own version, first seen Nov 15, 2026
- matthewdonsemail-lab/logtheir own version, first seen Nov 15, 2026
- cofin/flowtheir own version, first seen Nov 15, 2026
- baodhg/FLOWtheir own version, first seen Nov 8, 2026
Used with
Projects with it also use grill, status and 3 more.
Other files people use alongside this one.
Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
renatorozasSkillgrill
Adversarial code review. Don't let me ship until the changes pass your scrutiny.
Also in 3 projects that use this
AnthropicSkillstatus
Case status summary by audience — client-facing (plain language), internal (for the professor), or court-ready (formal caption format per local rules). Same facts, different framing and depth. Use when a student needs to update the client, brief the professor, or prepare a court status report.
Also in 3 projects that use this
BlockSkillplan
Draft a TDD, PRD, or lightweight plan in the conversation without committing it.
Also in 3 projects that use this
AWSSkillship
Finalize dev work on an issue branch — update docs, push, and open a PR.
Also in 3 projects that use this
GoogleSkilltdd
Test-driven development. Use when the user wants to build features or fix bugs test-first, mentions "red-green-refactor", or wants integration tests.
Also in 3 projects that use this
Other versions34 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 34
- react/react★ 253ksame name
- facebook/react★ 253ksame name
- software-mansion/react-native-audio-api★ 856same name
- simota/agent-skills★ 99same name
- panayiotism/claude-harness★ 79same name
- matthewdonsemail-lab/log★ 63same name
- rainday/smart-blog-skills★ 35same name
- cofin/flow★ 20same name
- basezh/agent-skills-on-base★ 15same name
- manhquydev/flow-skill★ 13same name
24 more not listed
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.
The file
We split this skill into the rules it gives an agent
2 rules. 0 are common, 1 is distinctive, and 1 is very specific to this project.
0
Common
Most of the 28 edited versions we read keep these.
1
Distinctive
This author's own choices. Worth a look.
1
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 10
Use this skill only when the user explicitly invokes it.
- Very specific: line 27
If the invocation includes --skip-rev (or the user otherwise says to skip review, e.g.
“Common” is measured: kept by at least 6 in 10 of the 28 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Flow
Use this skill only when the user explicitly invokes it.
Run the task through the pipeline: recon → spec → implement → review. Run each step in its own subagent (the Agent tool) instead of inline — that step's exploration and reasoning then stay out of this conversation's context, and only short summaries (or a genuine question) come back here. flow's job is to orchestrate, not to duplicate any step's own logic.
Running a step
For each step, in order:
- Tell the user, in one short line, which step is starting (e.g. "Spouštím recon…"). This is the only visibility they get while a step's subagent is working, since its own reasoning stays isolated from this conversation.
- Spawn a subagent with
Agent(general-purposetype;run_in_background: false, since the next step depends on this one finishing and there's nothing else useful to do meanwhile). Give it a self-contained prompt: the task description, which step it's running, that it's running as aflowsubagent (so it follows that skill's "running as a flow subagent" branch instead of trying to ask you directly), and to call the Skill tool for that step's name to get its actual instructions. Don't restate that skill's instructions yourself — the subagent loads them on its own. Set theAgentcall'sdescriptionto exactlyflow: <step>(e.g.flow: recon,flow: implement) — dcc's session status hook reads it to show the pipeline's progress indcc ps/dcc top. - Classify what comes back:
- Needs input — the entire output starts with
NEEDS INPUT:. Relay the question(s) to the user yourself — verbatim, or viaAskUserQuestionwhen they're a multiple-choice round — and wait for the answer. Once you have it, resume the same subagent withSendMessage(addressed to its agent id) carrying the answer, and re-classify its next output the same way — a step may need several rounds before it's actually done. - Failed — the subagent errored, stalled, or returned something that's neither a finished result nor a well-formed
NEEDS INPUT:block. Don't guess your way past this or invent a next step. Stop the pipeline, tell the user which step failed and what the subagent actually returned (or the error), and ask whether to retry that step, adjust the task, or stop here. - Finished — anything else. Tell the user in one short line that the step is done, then pull out only what the next step actually needs (a file path, a one-line summary) and carry that forward as input to its prompt — don't keep a finished subagent's full report in your own context beyond that; that defeats the point of running it as a subagent.
- Needs input — the entire output starts with
Skipping the review step
If the invocation includes --skip-rev (or the user otherwise says to skip review, e.g. "bez review", "no review this time"), run only recon → spec → implement and stop there — don't spawn a review step.
Result
After the last step that actually ran, give one short wrap-up: whether recon was reused or refreshed, the spec file written, what implement changed, and (if it ran) review's findings and whether they were addressed. Don't repeat each step's full output verbatim — that's exactly what running them as subagents was for.
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.

