| name | pacto-exec |
| description | Agent contract for the Pacto exec workflow. |
Pacto Exec Skill
Use this skill as an agent contract for the exec workflow in Pacto projects.
Objective
Execute plan tasks and register execution evidence in plan artifacts.
When To Use
Use after moving a plan to current to advance tasks while keeping execution evidence in plan documents.
Input Contract
Required Inputs
<state> and <slug> identifying an existing plan slice (state must be current).
Optional Inputs
--root <path> to target a specific project root.
--step <phase.task> to complete a specific task (for example, 1.2).
--note, --blocker, --evidence to append execution context.
--dry-run to preview updates without writing files.
Execution Contract
- Tool target: Cursor Agent and Codex (project skills under .agents/skills/)
- Recommended command: pacto exec [--root ] [--step <phase.task>] [--note ] [--blocker ] [--evidence ] [--dry-run]
Output Contract
- Marks pending tasks as completed in plan markdown checklists.
- Appends execution notes, blockers, and evidence references.
- Writes only plan artifact files (no source-code edits).
Validation Checklist
- Confirm target plan exists before execution updates.
- Confirm intended task selection (
next pending vs explicit --step).
- When dry-run is used, confirm no files are modified.
Failure Modes and Handling
- Missing/invalid plan target or invalid step id.
- No matching pending task for selected step.
- Filesystem write errors updating plan markdown.
Implementation Status
- Status: Implemented
- Fallback: Use
pacto status to inspect blockers or task state if execution updates cannot be applied.