# Grp06 Useful Codex Skills Useful Codex Skills > Implement ExecPlan - Skill: `tomevault-io/grp06-useful-codex-skills-useful-codex-skills` (Agent Skill, multi-file: 2 files) - Install (CLI): `npx skillmds@latest add tomevault-io/grp06-useful-codex-skills-useful-codex-skills` - Raw SKILL.md: https://api.skillmd.com/api/skills/tomevault-io/grp06-useful-codex-skills-useful-codex-skills/raw - Safety review: pending (external: skill-scanner PASS, skillspector PASS) - Works with: Claude Code, Claude.ai, OpenAI Codex - Category: Coding & Dev Tools - Author: tomevault-io (https://skillmd.com/u/tomevault-io) - Updated: 2026-09-17 - Page: https://skillmd.com/skills/tomevault-io/grp06-useful-codex-skills-useful-codex-skills --- # Implement ExecPlan Use the ExecPlan as the implementation contract for delivering both the intended behavior and the intended simplification of the system. ## Preferred Input Resolution Preferred target resolution order: 1. explicit work-item path supplied by the user 2. explicit `execplan.md` path supplied by the user 3. `.agent/active` when it points to a work item with: - `stage="plan"` and `state="completed"`, or - `stage="implementation"` and `state="blocked"` 4. the most recently updated work item under `.agent/work/` matching those same rules 5. legacy fallback: `.agent/execplan-pending.md` If no supported plan exists, stop and tell the user. ## Work Item Responsibilities If operating on a work item, read: - `meta.json` - `decision.md` when present - `execplan.md` This skill owns execution-state transitions in `meta.json`: - when starting work: `stage="implementation"`, `state="active"` - when blocked or only partially complete: `stage="implementation"`, `state="blocked"` - when implementation completes: `stage="implementation"`, `state="completed"` Do not rename plans or move directories to represent lifecycle state. ## Ousterhout Lens When the plan leaves room for judgment: - prefer deep modules over shallow wrappers - prefer interfaces that hide sequencing and policy - prefer fewer concepts, fewer knobs, and fewer special cases - prefer simple mental models over clever decomposition ## Workflow 1. Read the ExecPlan in full, then read `.agent/PLANS.md`. 2. Reconstruct the intended behavior, target boundary, and complexity dividend. 3. If available, read `decision.md` so the implementation stays aligned with the original decision rationale. 4. Before coding, update `meta.json` to `stage="implementation"` and `state="active"`. 5. Inspect the relevant code paths before editing. 6. Implement milestone by milestone. 7. Keep the ExecPlan's `Progress`, `Surprises & Discoveries`, `Decision Log`, and `Outcomes & Retrospective` sections up to date as you go. 8. After each meaningful slice, run the plan's validation steps or the nearest targeted verification. 9. If you finish the implementation, set `state="completed"`. 10. If you cannot finish safely in the current turn, record the blocker in the plan, set `state="blocked"`, and stop cleanly. ## Implementation-First Rule This skill is for execution, not more planning. You may update the ExecPlan during implementation only to: - record progress actually made - record discoveries from code already inspected - tighten milestones or acceptance criteria to match current repo reality - document design decisions that unblock the next implementation step Do not turn an implementation turn into another plan-improvement pass. ## Anti-Patterns - satisfying the letter of the plan while preserving the same interface burden - adding wrappers, adapters, or helper layers that hide little - pushing sequencing or policy outward to callers - finishing with only ExecPlan edits when implementation work remains - using file moves or renames as the primary state transition --- > Source: [grp06/useful-codex-skills](https://github.com/grp06/useful-codex-skills) — distributed by [TomeVault](https://tomevault.io).