generator-draft-contract-parallel

Stars13

Forks1

UpdatedMay 2, 2026 at 07:03

Internal Auto-Harness generator skill for parallel sprint contract drafting. Use only inside the Generator subagent when it is producing or revising the current parallel sprint contract.

Installation

Install with Codex or Claude Copy this prompt, paste it into Codex, Claude, or another assistant, and let it review the skill page and install it for you.

Source

redker56

redker56/auto-harness

View GitHub Repository View Creator Repositories

Download

Related occupationsSOC

Based on SOC occupation classification

Paralegals and Legal AssistantsLegal Occupations·SOC 23-2011

File Explorer

3 files

SKILL.md

readonly

name	generator-draft-contract-parallel
description	Internal Auto-Harness generator skill for parallel sprint contract drafting. Use only inside the Generator subagent when it is producing or revising the current parallel sprint contract.
user-invocable	false

Generator Draft Contract Parallel

This skill governs Generator contract mode.

At the start of this action, read these harness artifacts from the project:

.harness-parallel/status.md
.harness-parallel/intake.md
.harness-parallel/spec.md
.harness-parallel/design-direction.md
.harness-parallel/contracts/sprint-XX-contract.md when revising the current sprint contract.
.harness-parallel/contracts/sprint-XX-review.md when a prior review exists for the current sprint.

Then inspect the current project implementation relevant to this action:

Read the current implementation, architecture, routes, components, services, tests, and config that define the current sprint boundary and feasibility.
Use the codebase plus the harness artifacts to draft an implementation-ready contract.

Then read these skill references:

references/protocols/file-ownership.md
references/templates/sprint-contract.md

Follow these rules:

Treat .harness-parallel/intake.md, .harness-parallel/spec.md, the current sprint contract, and any current-sprint review artifact as authoritative inputs.
Write only the current sprint contract artifact.
If a review file already exists for the sprint, revise the contract directly against it.
Do not move the sprint toward implementation while the current sprint review result remains REVISE.
Treat an approved review as implementation guidance and cautions, not just a pass/fail stamp.
Keep the contract implementation-ready and evaluable from runtime plus named files.
Include both a human-readable Parallel Workstreams table and a machine-readable Dependency Graph JSON block.
The JSON graph is the source of truth for worktree scheduling.
Every node must declare non-overlapping owned_paths, explicit depends_on, and the behavior_ids it owns.
If the sprint should stay serial, emit exactly one node and say why in notes instead of omitting the graph.
Do not quietly violate locked architecture or stack choices; surface unavoidable deviations explicitly in the contract.
Do not modify status, review, QA, retest, or final artifacts.

More from this repository

same repository

evaluator-review-contract-parallel

redker56/auto-harness

Internal Auto-Harness evaluator skill for parallel sprint contract review before implementation. Use only inside the Evaluator subagent during review mode.

2026-05-0213

evaluator-write-final-parallel

redker56/auto-harness

Internal Auto-Harness evaluator skill for parallel final QA report aggregation. Use only inside the Evaluator subagent during evaluator_final_parallel.

2026-05-0213

evaluator-write-qa-parallel

redker56/auto-harness

Internal Auto-Harness evaluator skill for parallel sprint QA and QA report writing. Use only inside the Evaluator subagent during evaluator_qa_parallel.

2026-05-0213

evaluator-write-retest-parallel

redker56/auto-harness

Internal Auto-Harness evaluator skill for parallel sprint retest and retest report writing. Use only inside the Evaluator subagent during evaluator_retest_parallel.

2026-05-0213

generator-build-integrator

redker56/auto-harness

Internal Auto-Harness generator skill for worktree-merge build integration. Use only inside the parallel Generator integrator during build mode.

2026-05-0213

generator-build-worker

redker56/auto-harness

Internal Auto-Harness generator skill for one build dependency-graph node in a dedicated worktree.

2026-05-0213

name	generator-draft-contract-parallel
description	Internal Auto-Harness generator skill for parallel sprint contract drafting. Use only inside the Generator subagent when it is producing or revising the current parallel sprint contract.
user-invocable	false

Generator Draft Contract Parallel

This skill governs Generator contract mode.

At the start of this action, read these harness artifacts from the project:

.harness-parallel/status.md
.harness-parallel/intake.md
.harness-parallel/spec.md
.harness-parallel/design-direction.md
.harness-parallel/contracts/sprint-XX-contract.md when revising the current sprint contract.
.harness-parallel/contracts/sprint-XX-review.md when a prior review exists for the current sprint.

Then inspect the current project implementation relevant to this action:

Read the current implementation, architecture, routes, components, services, tests, and config that define the current sprint boundary and feasibility.
Use the codebase plus the harness artifacts to draft an implementation-ready contract.

Then read these skill references:

references/protocols/file-ownership.md
references/templates/sprint-contract.md

Follow these rules:

Treat .harness-parallel/intake.md, .harness-parallel/spec.md, the current sprint contract, and any current-sprint review artifact as authoritative inputs.
Write only the current sprint contract artifact.
If a review file already exists for the sprint, revise the contract directly against it.
Do not move the sprint toward implementation while the current sprint review result remains REVISE.
Treat an approved review as implementation guidance and cautions, not just a pass/fail stamp.
Keep the contract implementation-ready and evaluable from runtime plus named files.
Include both a human-readable Parallel Workstreams table and a machine-readable Dependency Graph JSON block.
The JSON graph is the source of truth for worktree scheduling.
Every node must declare non-overlapping owned_paths, explicit depends_on, and the behavior_ids it owns.
If the sprint should stay serial, emit exactly one node and say why in notes instead of omitting the graph.
Do not quietly violate locked architecture or stack choices; surface unavoidable deviations explicitly in the contract.
Do not modify status, review, QA, retest, or final artifacts.