agentleFS
Yash-SukhdeveYash-Sukhdeve/universal-workflow-system/.claude/skills/uws-orchestrate/SKILL.md

uws-orchestrate

Skill0 starsChanged 46 days agoMITView on GitHub

Run the current UWS phase as an autonomous, checkpoint-gated step — dispatch the phase's work to the matching subagent, then route its artifact through the human review gate. USE WHEN the user asks the framework to produce the current phase's deliverables, advance a UWS task, or "work on the current phase".

The file

We split this skill into the rules it gives an agent

6 rules. About 0 are common, 2 are distinctive, and 4 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈2

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 4

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Prepare the dispatch: Run: ./scripts/orchestrate.sh dispatch "<one-line task>" [target-rel-path] Parse the emitted DISPATCH: line — it …

    line 13
  2. Very specific:

    Read the brief: Read the brief file. It contains the goal, phase, target artifact path, and the deliverables checklist that will gate …

    line 17
  3. Distinctive:

    Do not let the subagent advance the workflow, checkpoint, or mark deliverables.

    line 23
  4. Very specific:

    Stage for review: When the artifact exists, run: ./scripts/orchestrate.sh collect "<summary>" This calls submit.sh to create a change …

    line 25
  5. Very specific:

    If no phase is active, tell the user to run ./scripts/sdlc.sh start (or research.sh start) and declare a goal with ./scripts/sdlc.sh goal …

    line 39
  6. Distinctive:

    Never write application code during a planning-phase (requirements/design) dispatch — those produce documents only.

    line 40

≈ “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.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

UWS Orchestrate

Drive one UWS phase end-to-end using real subagents, keeping the human review + deliverable gate in the loop. This is the execution core that binds the UWS state spine to Claude Code's Agent/Workflow primitives.

Procedure

  1. Prepare the dispatch. Run: ./scripts/orchestrate.sh dispatch "<one-line task>" [target-rel-path] Parse the emitted DISPATCH: line — it names agent, subagent (a .claude/agents/uws-<role>.md), phase, brief (workspace/<role>/TASK.md), and out (the artifact path to write).
  2. Read the brief. Read the brief file. It contains the goal, phase, target artifact path, and the deliverables checklist that will gate advancement.
  3. Run the subagent. Launch the named uws-<role> subagent (Agent tool; or Workflow for multi-agent phases) with a prompt that:
    • points it at the brief,
    • tells it to write its artifact to the out path (under workspace/<role>/, mirroring the intended repo path),
    • requires it to address every deliverable and STOP at its persona Quality Gate. Do not let the subagent advance the workflow, checkpoint, or mark deliverables.
  4. Stage for review. When the artifact exists, run: ./scripts/orchestrate.sh collect "<summary>" This calls submit.sh to create a change request under .uws/crs/ and post it to NOTIFICATIONS.md.
  5. STOP at the gate. Tell the user the change request is ready and show how to approve: ./scripts/review.sh approve <CR-ID> Do not approve on the user's behalf — approval is the human checkpoint (this is checkpoint-gated autonomy).
  6. After the user approves, mark the satisfied deliverables and advance: ./scripts/<methodology>.sh check <n> for each, then ./scripts/<methodology>.sh next. The hard gate will refuse to advance until every deliverable for the phase is checked (unless --force).

Notes

  • Methodology + phase + agent are resolved from .workflow/state.yaml; run ./scripts/orchestrate.sh status to see them.
  • If no phase is active, tell the user to run ./scripts/sdlc.sh start (or research.sh start) and declare a goal with ./scripts/sdlc.sh goal "<objective>" first — the deliverable gate only activates once a goal is declared.
  • Never write application code during a planning-phase (requirements/design) dispatch — those produce documents only.

Download now (ZIP)

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
What did you use it for? e.g. Reviewing pull requests in a Rails app

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.