| name | ticket-to-human-review |
| description | Use when an implementation contract needs human signoff before development/testing begins. This verifies the contract, not the implementation. |
Ticket To Human Review
Use this skill when a compiled implementation contract is ready for human
signoff.
This skill does not develop, test, merge, deploy, or release the work. It
answers one question: is this implementation contract clear, bounded, and
approved enough for the direct dev/testing composition to start?
Goal
Turn an implementation contract into one explicit human decision:
implementation contract
-> human contract review
-> SIGNED_OFF / CHANGES_REQUESTED / HOLD
-> direct iterative dev/testing composition, if signed off
SIGNED_OFF means ready for development and testing. It does not mean the work
is implemented, verified, merged, deployed, or released.
Inputs To Gather First
- implementation contract issue or packet
- active Implementation Contract parent
- source Design Contract refs
- source readiness/prep row, when one exists
- browser receipts, screenshots, annotations, or review evidence
- user-relief sentence and done signal
- primary implementation seam
- explicit non-goals
- SDLC composition map
- proof matrix and DB/Electron proof requirements
- governance grant, when mission autonomy or external actions are involved
- dependencies, blockers, and linked backend/design gaps
Start by reading:
- the implementation contract packet
../implementation-contract-conveyor/SKILL.md
../design-contract-conveyor/SKILL.md
.codex/skills/multi-agent-workflow/SKILL.md
.codex/skills/layered-verification/SKILL.md
Review Posture
The human is reviewing the contract, not code.
The operator should make the human decision narrow:
- what design truth was approved
- what one seam will be built
- what the contract explicitly excludes
- what dev/testing composition will run after signoff
- what proof is required before the work can later be called verified
- what governance grant the agent needs before an autonomous run
Do not ask the human to inspect code, run tests, choose worktrees, or debug
implementation details at this gate.
Workflow
1. Assert contract readiness
Do not start human signoff unless the contract includes:
- source Design Contract refs
- implementation parent and proposed child identity
- one primary implementation seam
- explicit non-goals
- source-of-truth and local-state boundary
- SDLC composition map
- proof matrix
- review and patch-loop expectations
- deployment/release boundary
If any of those are missing, return HOLD with the missing contract sections.
2. Check the design-to-implementation trace
Confirm the contract traces cleanly from approved design truth:
- approved design rules are present
- rejected, cut, or parked ideas did not sneak back into scope
- screenshots or browser receipts are referenced where useful
- implementation scope does not exceed the approved design cluster
If the contract expands beyond approved design truth, return
CHANGES_REQUESTED.
3. Check the dev/testing composition map
Confirm the contract names the iterative execution loops that will run after
signoff:
- lane fan-out through
$multi-agent-workflow
- non-author review swarm
- patch and re-review loop until zero blocking findings
$layered-verification loop until green or true blocker
- UI/Electron verifier lane when required
- Computer Use focus preflight for UI/Electron proof: exact Electron app path, raised/focused window, and real renderer HTML content in the accessibility tree before clicks or visual judgment
- persisted/DB-backed proof row when required
- receipt destinations
This is intentionally not a linear checklist. The contract should make it clear
where the system will spend more tokens/compute to raise quality.
4. Check human attention and governance
Confirm the contract says what the human is approving:
- what work may start
- what capabilities/data/actions the agent will need
- which approval boundaries apply
- what remains outside this implementation contract
- whether deployment/release is excluded and handled separately
If the agent grant is too vague for an autonomous run, return
CHANGES_REQUESTED.
5. Record the verdict
Allowed outcomes:
SIGNED_OFF: contract may enter direct dev/testing composition.
CHANGES_REQUESTED: contract needs edits before development/testing starts.
HOLD: contract cannot be judged because required context is missing or
upstream approval is unresolved.
For every outcome, record:
- reviewer
- timestamp
- implementation contract id or packet path
- verdict
- reason
- required edits, if any
- next workflow step
Output
Return:
- implementation contract reviewed
- verdict:
SIGNED_OFF, CHANGES_REQUESTED, or HOLD
- concise reason
- required contract edits, if any
- approved direct dev/testing composition, if signed off
- reminder that deployment/release is separate
Close Rule
This skill is complete when:
- the implementation contract was reviewed against approved design truth
- the dev/testing composition map was checked
- the governance and human-attention boundary was checked
- one explicit verdict was recorded
- no implementation, test, merge, deploy, or release work was performed inside
this skill