| name | Phase Prompting |
| description | Turn verified state into execution-ready prompts with wave structure, hard rules, acceptance criteria, and honest capability labels. |
Phase Prompting
Use this skill when the task is to write or revise:
- a phase prompt
- a round prompt
- a multi-wave execution packet
- a sustained coordination plan for executors
- a follow-on phase after reviewing delivery summaries or gate records
Core Principle
Write from current truth, not from assumptions. Start from what is already verified, distinguish what is only inferred, call out overclaimed, and identify what still blocks delivery.
Standard Prompt Shape
Follow this structure:
- Phase / round title
- Identity and mission
- First principles
- Current real state
- Key gaps
- Must-read files
- Hard rules
- Wave plan with estimated time
- Wave-by-wave tasks
- Required tests and build checks
- Verification assets / gate records
- Out-of-scope list
- Final output format
- Commit message
Operating Rules
1. Write from Current Truth
Start from:
- what is already verified
- what is only inferred
- what is overclaimed
- what still blocks delivery
Do not write prompts that assume a summary is correct when code or live behavior says otherwise.
2. Prefer One Execution Round Over One Giant Roadmap
The actual prompt handed to an executor should usually be one round:
- about
2-3 hours
- about
5-7 waves
- with one dominant delivery goal
- with explicit branch conditions when the round can split
If the next step depends on whether A passes or fails, write Round 2A and Round 2B instead of one oversized prompt.
3. Every Round Must Create Real User Value
Do not write rounds that mostly:
- rename things
- restate earlier work
- add surface labels without runtime truth
- repeat already-proven chains
Prefer rounds that close one real gap:
- a path becomes truly runnable
- a false capability claim becomes honest and user-visible
- a long-task path becomes truly deliverable
- a benchmark or release gate becomes real
4. Force Multi-Layer Progress
Unless explicitly narrowed, each round should try to advance:
- backend or orchestration behavior
- at least one user entry surface
- tests
- docs or protocol or verification assets
5. Capability Honesty is Mandatory
Require prompts to distinguish:
supported
partial
best_effort
manual-only
blocked
unsupported
Never let a prompt reward inflated claims.
6. Runtime Safety Must Be Explicit
For heavy operations, long context, or risky changes, prompts should include:
- resource guard expectations
- isolated runtime rules when needed
- laddered verification instead of blind escalation
- blocked outcomes as valid outcomes
7. Prompt Archiving is the Default
When producing an execution-ready prompt:
- Save the final prompt into the project repo unless explicitly told not to
- Archive prompts under a project-specific directory
- Use deterministic filenames:
phase-XX-round-Y.md, phase-XX-round-Ya.md, phase-XX-round-Y-revision.md
- Archive the final prompt only, not chain-of-thought
Wave-Writing Rules
For each wave, include:
- goal
- why this wave matters
- concrete tasks
- acceptance criteria
Good waves:
- fix a real broken route
- replace an overclaim with a real verified path
- turn a doc concept into a real gate
- sync API, UI, protocol, and tests
Weak waves:
- "review some files"
- "polish docs"
- "clean up naming"
- "do more validation" without a delivery objective
Branching Rules
When the round depends on a high-risk result, write the branch into the prompt:
- if primary path passes, proceed to productize it
- if it fails, fall back to alternative approach
- if blockers exist, mark blocked and close honesty first
Do not make later waves depend on a success path that may never happen.
Required Final Report Sections
Every prompt should require a final report with:
Modified files
Wave-by-wave outcomes
New user-visible capabilities
New/updated APIs
Tests run
Build/check run
Release-readiness delta
Remaining blockers before delivery
Why this phase materially advances toward delivery
Output Expectation
When using this skill, produce one of:
- a complete execution-ready prompt
- a split plan: roadmap plus
Round 1 prompt plus branch prompts
- a revision of an earlier phase prompt that tightens truth, scope, and gates
The output should be ready to hand to an executor with minimal editing.