Create Execution Checklist.prompt
---
id: "02"
phase: 2
name: "Create execution checklist"
template: "from-adr"
targets:
- "app"
commit: "docs: add execution checklist"
references:
- "docs/adr-001-runtime-boundaries.md"
- "docs/adr-002-recovery-contract.md"
required_reading:
- "docs/adr-001-runtime-boundaries.md"
- "docs/adr-002-recovery-contract.md"
context_files:
- "workspace/RUNTIME_BOUNDARIES.md"
depends_on:
- "01"
verify:
files_exist:
- "EXECUTION_CHECKLIST.md"
contains:
- path: "EXECUTION_CHECKLIST.md"
text: "Define verifier-owned completion contracts."
- path: "EXECUTION_CHECKLIST.md"
text: "Attach required reading to each prompt."
changed_paths_only:
- "EXECUTION_CHECKLIST.md"
simulate:
attempts:
- messages:
- "Reading the ADRs plus the runtime boundary summary, then writing the execution checklist."
writes:
- path: "EXECUTION_CHECKLIST.md"
text: |-
# Execution Checklist
- Define verifier-owned completion contracts.
- Attach required reading to each prompt.
- Use retries and repair based on verifier results.
---
# Create execution checklist
## Required Reading
- `docs/adr-001-runtime-boundaries.md`
- `docs/adr-002-recovery-contract.md`
- `workspace/RUNTIME_BOUNDARIES.md`
## Architecture Context
- The packet runtime needs clear execution boundaries.
- Recovery is driven by verifier results, not only provider success.
## Mission
Create `EXECUTION_CHECKLIST.md` in the target repo using the ADR decisions and
the runtime boundary summary from prompt `01`.
## Deliverables
- `EXECUTION_CHECKLIST.md` with actionable execution rules
## Non-Goals
- Do not modify any other files.
- Do not restate the ADRs verbatim.
## Verification Notes
- The file must exist.
- It must include the verifier-owned completion and required-reading lines.
- No other files should change.when to use it
Community prompt sourced from the open-source GitHub repo nshkrdotcom/prompt_runner_sdk (MIT). A "Create Execution Checklist.prompt" style prompt — adapt the placeholders and specifics to your task. Imported as-is and not independently retested here, so check the output before relying on it.
tags
codingcommunitydeveloper
source
nshkrdotcom/prompt_runner_sdk · MIT
more in Coding
Coding✓ tested
Senior code review (strict mode)
senior staff engineer running a merciless but fair review
Coding✓ tested
Debug by hypothesis, not by guessing
debugging partner who forms theories before touching code
Coding✓ tested
Generate tests from described behavior
test engineer who writes tests that would actually catch regressions