plan-execution-packetlisted

# Skill: Plan Execution Packet <!-- Knowledge source: GENERALIZE of the private EI's plan-execution-packet SKILL.md. Kept as a thin pointer to playbooks/execution-packet-planning.md (the canonical, detailed source) plus the skill-specific stop conditions - per this framework's D-007, one canonical source per workflow, generated adapter loaders only. --> Use when work can't be handed to an execution agent as one ad-hoc session: an idea/PRD/epic, a large piece of work, findings from review or testing, or cross-project-instance scope. Full workflow: [`playbooks/execution-packet-planning.md`](../../playbooks/execution-packet-planning.md). Templates: `templates/packet-*.md` and `templates/session-launch.md`. ## Process 1. Classify the work and confirm it actually needs a packet (see [`playbooks/session-preparation.md`](../../playbooks/session-preparation.md) Step 1) - otherwise use [`templates/task-scope.md`](../../templates/task-scope.md) instead. 2. Search `knowledge/` for prior related work, incidents, or rejected approaches (`scripts/eif_search_knowledge.py`). Record what was found. 3. Write the charter (problem, goal, explicit in/out of scope, boolean DoD) before anything else. 4. Write the facts file: verify current state against fresh source/tool state, not memory. Build the carry-over ledger - every relevant item from prior planning gets an explicit disposition. 5. Write the decisions file: resolve every point where an executing agent would oth