Agent Skills: Protocol Writer

|

UncategorizedID: willoscar/research-units-pipeline-skills/protocol-writer

Install this agent skill to your local

pnpm dlx add-skill https://github.com/WILLOSCAR/research-units-pipeline-skills/tree/HEAD/.codex/skills/protocol-writer

Skill Files

Browse the full folder contents for protocol-writer.

Download Skill

Loading file tree…

.codex/skills/protocol-writer/SKILL.md

Skill Metadata

Name
protocol-writer
Description
|

Protocol Writer

Transforms the review question into an executable evidence-review protocol.

Inputs

Required:

  • STATUS.md

Optional:

  • GOAL.md
  • DECISIONS.md
  • queries.md

Output

  • output/PROTOCOL.md

Contract

The protocol must contain:

  • review questions
  • source/search specification
  • inclusion and exclusion clauses with stable IDs
  • screening plan
  • extraction schema
  • bias plan
  • explicit HUMAN approval note

Script boundary

scripts/run.py should:

  • read current review context
  • materialize an operational protocol
  • write a stable extraction schema and clause IDs

It should not perform retrieval or screening itself.

Acceptance

  • output/PROTOCOL.md exists
  • includes extraction schema and protocol clause IDs
  • is operational enough for downstream screening

Non-goals

  • writing final synthesis prose
  • screening candidate papers