Implement a Bounded Code Change

Implement one named change and leave merge with the human owner. A diff is output. Correctness and beneficiary value remain unproved.

Use when. A named change, acceptance criteria, and a human owner are supplied.

Do not use when. Acceptance criteria, constraints, or the owner are missing, or you want the model to merge or release.

Your input. Repository context and exact revision; the named change; acceptance criteria; constraints; named owner.

Your quality check. The diff matches the named change. Invented dependencies and silent signature changes are called out. Merge stays blank.

Put this to work

Prepare one bounded code change

For the named repository owner

Copy this prompt. Paste it with the relevant source material into Claude, ChatGPT, or any AI assistant, then review the result before acting.

# Role and job
Help a named human owner prepare one Reviewable code change for Implement a Bounded Code Change.

# Outcome
Produce one Reviewable code change. Success means the result matches the supplied job, unknowns are visible, and the human decision remains blank.

# Inputs
- Supplied context: [paste]
- Constraints and permitted claims: [paste]
- Named owner: [name]

# Boundaries
You may draft, compare, and expose gaps. Do not invent proof, contact a person, publish, spend, merge, release, or make the human decision.

# Output
Return the Reviewable code change; assumptions; unknowns; risks; and the exact human decision required.

# Stop and recovery
Stop when the owner, the job boundary, or a required input is missing. Return the blocker and the smallest missing input.

# Final check
Verify the result stays inside the supplied inputs and the consequential decision remains with the named human owner.

Human decision — cannot be delegated. Merge, request a revision, or stop. AI may draft and explain only. No clinic CTA.

Outcome

A reviewable diff for one named change, with unknowns visible and merge left blank.

Use When

Use when a named change, acceptance criteria, and a human owner are supplied.

Inputs

Repository context and exact revision; the named change; acceptance criteria; constraints; named owner.

Recipe

Copy the prompt. Paste only supplied material. Leave the human decision blank.

Human and AI Control

An assistant may draft and report. Named repository owner controls merge, release, spending, and accepted risk. The tool is a player, not the job.

Output and Handoff

Return the Reviewable code change and the blank human decision.

Evidence and Stop

Demand is hypothesized and the outcome is untested. Stop when the owner, the job boundary, or a required input is missing.

Reliability

Skill and tool: bounded code change using a repository-authorised coding assistant. The Reviewable code change is the pattern; constraint binding and human merge is the protocol. Retain the returned brief as evidence. Accept only a result the named human can verify against the supplied inputs.

Next

Return the result to the named owner. Do not register autonomous execution until a completed prompt trial and an accepted execution contract exist.