> orchestrate-next-milestone
Implement the manager-prioritized milestone of an existing plan and carry it to closure, including adjacent fixes needed to pass gates and close the milestone cleanly. Use when a manager/reviewer says to start the next milestone and expects end-to-end delivery with evidence.
curl "https://skillshub.wtf/vosslab/vosslab-skills/orchestrate-next-milestone?format=md"Orchestrate Next Milestone
Overview
Execute the selected plan milestone to closure, not just partial checkboxes. Prioritize finishing the milestone with passing gates and clear evidence.
Terminology Alignment
- Milestone: primary planning unit with deliverables and gates.
- Workstream: parallel lane inside a milestone.
- Work package: coder-sized chunk with acceptance criteria and verification commands.
- Patch: reviewable code change set used in summaries and changelog entries.
- Stage / Step / Pass: durable pipeline terms for code identifiers when needed.
- If older plans use "phase", treat it as legacy milestone wording and do not introduce new phase terminology.
Operating mode
- Treat the manager-selected milestone as the primary target.
- Implement code, tests, and documentation updates required to close that milestone.
- Include adjacent fixes when they are necessary to pass milestone gates or prevent immediate regressions.
- Keep changes contract-driven and avoid case-specific hacks.
Inputs to read first
- Review STYLE guides in AGENTS.md, docs/PYTHON_STYLE.md, and docs/REPO_STYLE.md
- Manager directive naming the target milestone (or latest explicit manager instruction).
- Target plan in
docs/active_plans/(if present in the target repo; or use specified path). refactor_progress.mdfor active/completed context (if present in the target repo).- Related archive precedent if cited by the plan.
- Current repo state (
git status --short,git diff).
Mandatory constraints
- Treat the plan design philosophy near the top as binding architecture policy.
- Do not stop at partial progress when the remaining work is directly required to close the selected milestone.
- Do not ask for guidance before attempting concrete unblock steps.
- Do not weaken strict gates to force a pass.
- Do not move geometry/runtime policy into tool/report layers unless the plan explicitly requires it.
- Do not claim completion without reproducible command evidence.
Scope policy
- Primary scope: selected milestone deliverables and done checks.
- Allowed adjacent scope: fixes that are strictly required to satisfy selected-milestone gates.
- Deferred scope: improvements not required for selected-milestone closure.
When adjacent scope is used, record:
- why it was required,
- what gate it unblocked,
- why deferral would leave the milestone incomplete.
Workflow
- Lock the target milestone:
- Identify exact milestone id/name and extract deliverables, done checks, and gates.
- Record explicit out-of-scope items from later milestones.
- Build execution map:
- Map deliverables to file paths.
- Map work packages and done checks/gates to exact commands.
- Identify likely adjacent fixes needed for closure.
- Implement to closure:
- Apply required edits for the milestone.
- Apply adjacent unblock fixes when needed.
- Keep design-philosophy alignment explicit in decisions.
- Validate gates:
- Run milestone-required tests/commands first.
- Run broader regression gates required by the plan.
- Run targeted
pytestregression:source source_me.sh && python3 -m pytest -q tests/test*.py -k <file>/py. - Treat any failing
pytestrun asNot completeuntil fixed. - Iterate until gates pass or a concrete blocker remains.
- Prepare evidence:
- Capture command outputs and pass/fail results.
- Capture before/after artifacts when visuals/behavior are part of acceptance.
- Capture residual risks and ownership.
- Close out docs:
- Update
docs/CHANGELOG.mdwith concrete changes and validation commands/results. - Update plan status language only when evidence supports closure.
Implementation checklist template
- Milestone target:
- Deliverables completed:
- Adjacent fixes applied (and why):
- Deferred items:
- Files changed:
- Work packages completed:
- Patches delivered:
- Commands run:
- Gate results:
- Remaining blockers:
- Completion decision:
CompleteorNot complete
Failure handling
- If milestone target is ambiguous, pick the most recent manager-specified milestone and state the assumption.
- If gates conflict, follow the plan-defined gate hierarchy and report the conflict explicitly.
- If blocked, report exact blocker plus attempted mitigations; do not present partial work as complete.
Manager completion report (required)
Submit one manager-grade close-out report at milestone close-out using this structure.
Required sections:
Milestone decision:CompleteorNot complete.Scope execution: what was completed, what was deferred, and why.Files changed: exact paths grouped by runtime/tests/docs/tooling.Gate evidence: exact commands with PASS/FAIL.Design philosophy alignment: how implementation followed plan philosophy and avoided hacks.Visual/behavior evidence: before/after paths and measured values (when required).Patch summary:Patch 1,Patch 2, ... with intent and touched components.Open risks: unresolved risks with impact and owner.Next action: one prioritized next step.
Report template:
Milestone decision: Complete | Not complete
Scope execution:
- Completed:
- Deferred:
- Adjacent fixes (required for closure):
Files changed:
- Runtime:
- Tests:
- Docs:
- Tooling:
Patch summary:
- Patch 1:
- Patch 2:
Gate evidence:
- <command>: PASS/FAIL
- <command>: PASS/FAIL
Design philosophy alignment:
- <how implementation stayed aligned and avoided hacks>
Visual/behavior evidence:
- Before: <path>
- After: <path>
- Measurements: <metric=value, threshold=value>
Open risks:
- <risk, impact, owner>
Next action:
- <single prioritized next step>
Output contract
Return results in this order:
- Milestone completion decision (
CompleteorNot complete). - Deliverables implemented (mapped to files).
- Work packages and patch summary (
Patch 1,Patch 2, ...). - Validation evidence (commands and outcomes).
- Known gaps and risks.
- Next action recommendation.
> related_skills --same-repo
> webwork-writer
Create, edit, and lint WeBWorK PG/PGML questions following docs/webwork guidance, HTML whitelist constraints, and renderer-based lint checks. Use for tasks like authoring new PGML problems, adjusting randomization or grading, fixing PGML rendering issues, and running renderer API linting.
> unit-test-starter
Generate thorough Python 3 pytest unit tests across a repo by scanning every *.py file and each function, writing one test module per source file while skipping IO/network behavior and documenting gaps.
> skill-writing-guide
Guide for authoring Agent Skills (SKILL.md). Covers the open standard format, required frontmatter, directory layout, progressive disclosure, description writing, and best practices. Use when creating a new skill, improving an existing skill, or learning how skills work.
> readme-fix
Standardize README.md to match repo conventions. Brief purpose, quick start, and links to docs/. Keep content verifiable, concise, and ASCII. Use when README.md drifted or is missing key pointers.