human-checkpoint

Human Checkpoint (HITL sign-off)

Safety Notice

This listing is imported from skills.sh public index metadata. Review upstream SKILL.md and repository scripts before running.

Copy this and send it to your AI assistant to learn

Install skill "human-checkpoint" with this command: npx skills add willoscar/research-units-pipeline-skills/willoscar-research-units-pipeline-skills-human-checkpoint

Human Checkpoint (HITL sign-off)

Goal: make human approvals explicit and auditable, so downstream units can safely proceed.

This skill is intentionally simple: it standardizes how a human signs off in DECISIONS.md .

Inputs

Required:

  • DECISIONS.md

Optional:

  • UNITS.csv (to confirm which checkpoint is currently blocked, e.g., C1 /C2 )

  • STATUS.md (to confirm the current checkpoint)

Outputs

  • DECISIONS.md (updated checkbox + short sign-off note)

Procedure

Identify the blocked checkpoint

  • Check the runner message (or STATUS.md ), or inspect UNITS.csv for the first owner=HUMAN unit that is BLOCKED .

Open DECISIONS.md and find the approvals checklist

  • Look for a line like: - [ ] Approve C2 ...

Review the artifacts required by that checkpoint

  • The pipeline doc (pipelines/*.pipeline.md ) usually lists what to review (e.g., protocol, outline, module plan).

  • If the checkpoint block is missing, add a short checklist into DECISIONS.md (what you reviewed).

Approve

  • Tick the checkbox: - [x] Approve C*

  • Append a short note under the checkpoint block (recommended):

  • Date

  • Artifacts reviewed

  • What you approved + constraints (if any)

  • Signed by

Acceptance

  • The correct Approve C* checkbox in DECISIONS.md is ticked ([x] ).

  • Any added sign-off note does not silently expand scope or contradict the run goal.

Troubleshooting

The approvals checklist is missing in DECISIONS.md

Fix:

  • Run pipeline-router for the relevant checkpoint block, or add a minimal block manually:

Approvals (check to unblock)

  • Approve C2

Source Transparency

This detail page is rendered from real SKILL.md content. Trust labels are metadata-based hints, not a safety guarantee.

Related Skills

Related by shared tags or category signals.

Research

pdf-text-extractor

No summary provided by upstream source.

Repository SourceNeeds Review
Research

latex-compile-qa

No summary provided by upstream source.

Repository SourceNeeds Review
Research

draft-polisher

No summary provided by upstream source.

Repository SourceNeeds Review
Research

citation-verifier

No summary provided by upstream source.

Repository SourceNeeds Review