cglab-public/agenfk/agenfk-flow
Claude Code and agent skill by cglab-public/agenfk
- community
Checking that the SKILL.md file is still in the repository.
Interactively create or edit an AgenFK workflow flow in chat.
Install
npx vskill@latest install cglab-public/agenfk/agenfk-flowSKILL.md
AgenFK Flow Manager
This skill guides you through creating or editing a custom workflow flow for an AgenFK project. A flow defines the ordered steps (statuses) that items move through — replacing the default TODO → IN_PROGRESS → REVIEW → TEST → DONE pipeline with a tailored one for your team.
How to use this skill
Invoke this skill by running the /agenfk-flow slash command or by asking:
"Help me create a new flow" / "I want to set up a custom workflow"
Conversation Protocol
Follow these steps in order. Ask one section at a time — do not dump all questions at once. Use the Question UI (default_api:question) to collect each input interactively.
Step 1 — Identify the project
- Run
agenfk current-projectto get theprojectId(it resolves the nearest.agenfk/local.json, walking up from the cwd). - If not found, call
list_projects()via MCP and use the Question UI to ask which project. - Confirm: "I'll create this flow for project [name] (
[projectId]). Is that correct?"
Step 2 — Flow identity
Use Question UI to ask:
- Flow name (required, machine-safe, e.g.
security-review,ml-training): no spaces, lowercase-hyphenated recommended. - Description (optional): one sentence describing when to use this flow.
Step 3 — Collect steps
Explain to the user:
"A flow is a sequence of steps. Each step represents a status an item can be in.You need at least 2 steps. The last step is usually a terminal step (equivalent to DONE).Tell me about each step one at a time, or give me the full list at once."
For each step, collect via Question UI:
| Field | Required | Notes | |
|---|---|---|---|
name | Yes | Machine-safe identifier, e.g. IN_PROGRESS, QA, SHIPPED | |
label | No | Human-friendly display name (defaults to name) | |
exitCriteria | No | What must be true before leaving this step | |
isSpecial | No | true if this is a terminal/archive step (like DONE) |
Steps are automatically ordered in the sequence you provide them.
Example steps for a security-focused flow:
TODO— "Not started"IN_PROGRESS— "Being implemented"SEC_REVIEW— "Security review", exitCriteria: "No critical CVEs, signed off by security team"STAGING— "Deployed to staging", exitCriteria: "All integration tests pass on staging"DONE— "Released", isSpecial: true
Step 4 — Preview and confirm
Display the collected flow as a table:
Use Question UI to ask: "Does this look right? (yes / edit / cancel)"
Step 5 — Create the flow
Once confirmed, use the server REST API directly:
Or via CLI:
Display every MCP call parameter and return value (Opencode transparency requirement).
Step 6 — Optionally activate the flow for the project
Use Question UI: "Would you like to activate this flow for project [name] now? (yes/no)"
If yes:
Or via REST:
Step 7 — Summary
Report back:
- Flow ID and name
- Number of steps created
- Whether it was activated for the project
- CLI command to inspect it:
agenfk flow show <flowId>
Editing an existing flow
If the user wants to edit a flow instead of creating one:
- List flows:
agenfk flow listorGET /flows - Show the target flow:
agenfk flow show <id> - Run:
agenfk flow edit <id>— or usePUT /flows/<id>via REST with the full updated flow body. - Confirm changes.
Notes
- Flow names must be unique within a project.
- The gatekeeper returns the active flow's steps and the current step's exit criteria —
agenfk gatekeeperand theworkflow_gatekeeperMCP tool both report them, so all platforms benefit automatically once a flow is activated. - To reset a project back to the default flow:
agenfk flow reset --project <projectId> - To share a flow with the community:
agenfk flow publish <flowId>
Security
- Security scanPassed, 100 of 100Pattern checks on the skill's files.
- Model reviewNot runA model review is scheduled for the next rescan.
Quality evaluation
- Task evaluationNot runNo evaluation yet. Run one in Skill Studio.
Source
- SKILL.md in the repositoryCheckingskills/opencode/agenfk-flow/SKILL.md
Popularity
- Repository stars
- 67
- Repository forks
- 14
- 7-day trend score
- 21
Stars and forks are for the whole GitHub repository, not this skill alone. Not refreshed yet.
Works with
Any agent vskill installs to: 56 targets in all. These 39 have a page on this site.
- Claude Code
- Cursor
- GitHub Copilot
- Windsurf
- Codex
- Gemini CLI
- Cline
- Amp
- Roo Code
- Goose
- Continue
- Kilo Code
- Junie
- Trae
- OpenHands
- Qwen Code
- AI Agent
- Tabnine Chat
- Sourcegraph Cody
- Amazon Q
- Firebase Genkit
- Sweep
- Devin
- Phind
- Mentat
- Claude Desktop
- Aider
- GPT Pilot
- Plandex
- bolt.new
- v0
- Replit Agent
- Anse
- TypingMind
- BetterChatGPT
- LibreChat
- Warp AI
- Pieces
- Fig AI
Embed badge
[](https://verified-skill.com/skills/cglab-public/agenfk/agenfk-flow)