Execution Plans

This skill helps you produce execution-ready plans and maintain workdocs so complex work can be resumed, reviewed, and audited across sessions without relying on provider-specific features.

Purpose

• •Turn a vague request into an executable plan with explicit outputs, checks, dependencies, and gates.
• •Keep scenario-local task state in workdocs (plan.md, context.md, tasks.md) so progress survives session boundaries.
• •Reduce risk on non-trivial changes by making validation and human approvals explicit.

When to use

Use this skill when:

• •The task is multi-step (dependencies) or cross-scope (modules + system + ops + knowledge).
• •The task requires approval gates (security, registries, production-adjacent config).
• •You need durable handoff notes (context resets, multi-session work).

Do not use this skill for trivial one-shot edits; use a short micro plan in chat.

Core concepts (repository-native)

Workdocs are scenario-local state

• •Workdocs live under workdocs/ directories (typically workdocs/active/T-YYYYMMDD-slug/).
• •Workdocs are the source of truth for 鈥渨hat was planned, what changed, and what鈥檚 next鈥?

Execution plans must be verifiable

• •Each planned unit of work should name concrete outputs and checks.
• •Add gates whenever progress requires human input/approval.

Minimal workflow

1. •Create/open workdocs/active/T-YYYYMMDD-slug/.
2. •Author an execution-ready plan.md (YAML todos + narrative).
3. •Maintain context.md and tasks.md as you work.
4. •Before ending a session, run a handoff update pass to capture the exact state and next steps.

Related