agentleFS
joris887joris887/exosuit/.claude/skills/phase-review/SKILL.md

phase-review

Skill9 starsChanged 3 months agoMITView on GitHub

Phase transition review — walkthrough built features, validate assumptions, refresh research, decide direction, and plan next phase.

1 other project uses a version of this.

Adopted by

Most recently taken up by

The file

Show its rules and text

We split this skill into the rules it gives an agent

16 rules. About 0 are common, 8 are distinctive, and 8 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈8

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 8

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Read these files (skip any that don't exist with a warning):

    line 33
  2. Very specific:

    Read epic files from docs/reference/backlog/ to determine:

    line 45
  3. Very specific:

    PIVOT: Run /discover Phases 1-3 with new direction Save to vision/phase-N+1-discovery.md.

    line 70
  4. Very specific:

    Update docs/reference/BACKLOG_INDEX.md with new epics

    line 83
  5. Very specific:

    Create new epic files in docs/reference/backlog/

    line 84
  6. Very specific:

    Update docs/progress.md with phase review outcomes

    line 85
  7. Very specific:

    Update docs/reference/DECISION_LOG.md with any revised decisions

    line 86
  8. Very specific:

    Read scale from vision/classification.md:

    line 100
  9. Distinctive:

    Quick Build: Run only E0N-001 + E0N-004 + E0N-006 (3 stories)

    line 101
  10. Distinctive:

    Standard: Run all 6 stories

    line 102
  11. Distinctive:

    Platform: Run all 6 + Architecture Review (E0N-007)

    line 103
  12. Distinctive:

    Pioneering (post-spike): Run E0N-001 + E0N-004 + full /discover re-entry

    line 104
  13. Distinctive:

    NEVER skip the user walkthrough (E0N-001) — this is the core value of phase review

    line 108
  14. Distinctive:

    NEVER auto-decide PIVOT vs PERSEVERE — this is always a user decision

    line 109
  15. Distinctive:

    ALWAYS include Phase Transition stories in the next phase backlog

    line 110
  16. Very specific:

    ALWAYS update ASSUMPTION_REGISTER and DECISION_LOG with review outcomes

    line 111

≈ “Common” is an estimate: at least 1 in 5 of the 372,429 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

phase-review

Running Phase Transition Review for phase: $ARGUMENTS

Process Flow

START → Load context (ASSUMPTION_REGISTER, DECISION_LOG, BACKLOG_INDEX, progress.md)
  → Identify completed epics
  → Walk through Phase Transition Stories (E0N-001 → E0N-006) interactively
  → Generate outputs to docs/reviews/ and vision/
  → If user approves next phase: invoke /ideate with new vision
  → DONE

1. Load Context

Read these files (skip any that don't exist with a warning):

  • docs/reference/ASSUMPTION_REGISTER.md — assumptions to validate
  • docs/reference/DECISION_LOG.md — decisions to review
  • docs/reference/BACKLOG_INDEX.md — epic completion status
  • docs/progress.md — sprint history and metrics
  • vision/project-pitch.md — original vision and No-Gos
  • vision/classification.md — archetype and scale

Determine which phase we're reviewing from $ARGUMENTS or by counting completed review epics.

2. Identify Completed Work

Read epic files from docs/reference/backlog/ to determine:

  • Which epics are complete (all stories done)?
  • Which stories were carried over?
  • What was the overall delivery rate?

3. Execute Phase Transition Stories

Walk through each story interactively. Read .claude/skills/discover/references/phase-transition-template.md for the full story definitions.

3A. Feature Walkthrough (E0N-001)

Walk through every built feature with the user. Copy template from .claude/skills/discover/assets/phase-walkthrough.md to docs/reviews/phase-N-walkthrough.md and fill it in.

3B. Assumption Validation (E0N-002)

Go through ASSUMPTION_REGISTER.md. For each HIGH-impact assumption: VALIDATED, INVALIDATED, or STILL UNKNOWN? Update the register.

3C. Research Refresh (E0N-003)

Load archetype from vision/classification.md. Run archetype-specific research from .claude/skills/discover/references/research-protocols.md (RC6 checkpoint). Save to docs/research/phase-N-refresh.md.

3D. Pivot or Persevere (E0N-004)

Load success criteria from .claude/skills/discover/references/engineering-by-archetype.md. Present current status vs targets. Apply decision framework (SCALE / PERSEVERE / PIVOT / KILL). Save to docs/reviews/phase-N-direction.md.

3E. Next Phase Elicitation (E0N-005)

Based on direction from 3D:

  • SCALE: Elicit growth features and optimization targets
  • PERSEVERE: Identify validation experiments for remaining unknowns
  • PIVOT: Run /discover Phases 1-3 with new direction Save to vision/phase-N+1-discovery.md.

3F. Next Phase Backlog (E0N-006)

Generate next batch of epics from the phase N+1 discovery. The last epic is AGAIN a Phase Transition epic — maintaining the infinite cycle.

<HARD-GATE> User approves the next-phase backlog before writing to files. </HARD-GATE>

4. Generate Outputs

After user approval:

  • Update docs/reference/BACKLOG_INDEX.md with new epics
  • Create new epic files in docs/reference/backlog/
  • Update docs/progress.md with phase review outcomes
  • Update docs/reference/DECISION_LOG.md with any revised decisions

Graceful Degradation

DependencyIf Missing
ASSUMPTION_REGISTER.mdSkip assumption validation (E0N-002), note gap
DECISION_LOG.mdSkip decision review, note gap
vision/classification.mdAsk user for archetype during research refresh
WebSearchSkip research refresh (E0N-003), note "no internet"
progress.mdSkip metrics comparison in pivot-or-persevere

Scale Adaptations

Read scale from vision/classification.md:

  • Quick Build: Run only E0N-001 + E0N-004 + E0N-006 (3 stories)
  • Standard: Run all 6 stories
  • Platform: Run all 6 + Architecture Review (E0N-007)
  • Pioneering (post-spike): Run E0N-001 + E0N-004 + full /discover re-entry

Rules

  • NEVER skip the user walkthrough (E0N-001) — this is the core value of phase review
  • NEVER auto-decide PIVOT vs PERSEVERE — this is always a user decision
  • ALWAYS include Phase Transition stories in the next phase backlog
  • ALWAYS update ASSUMPTION_REGISTER and DECISION_LOG with review outcomes

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.