cj-implement-from-spec

name: CJ_implement-from-spec description: "Implement a CJ_personal-workflow work-item from its input artifacts. Reads per-type spec (SPEC+DESIGN for user-stories, RCA+test-plan for defects, TRACKER+test-plan for tasks; features delegate to a child user-story via AUQ), plans against Components Affected / Data Flow, writes code via Read/Edit/Write. Sensitive-surface AUQ for catalog/manifest/validator edits; propose-and-confirm by default (--auto for trivial ≤2-file changes). Idempotent. Use when: 'implement this work-item', 'write the code for this spec'." version: 1.0.0 allowed-tools: - Bash - Read - Write - Edit - Glob - Grep - AskUserQuestion

Preamble

Check for collection updates (silent if none, banner if a newer version is available):

_UC="${CJ_SHARED_SCRIPTS:-$HOME/.claude/_cj-shared/scripts}/skills-update-check"
[ -x "$_UC" ] && "$_UC" 2>/dev/null || true

Verify this is a git repository:

git rev-parse --show-toplevel 2>/dev/null || echo "NOT_A_GIT_REPO"

If NOT_A_GIT_REPO: tell the user "Error: /CJ_implement-from-spec requires a git repository." and stop.

Update Nudge Handling (skip silently if preamble printed nothing about updates)

Same as /CJ_personal-workflow: if preamble output contains SKILLS_UPGRADE_AVAILABLE <old> <new>, follow the upgrade flow defined in ~/.claude/skills/CJ_personal-workflow/SKILL.md. If SKILLS_JUST_UPGRADED <from> <to>, print "claude-skills-templates upgraded to v<to> (was v<from>)" and continue.

Path Resolution

Resolve skill assets using a 2-level fallback chain. This skill depends on CJ_personal-workflow's manifest + WORKFLOW.md as its runtime source of truth (boundary check via /CJ_personal-workflow check).

_REPO_ROOT=$(git rev-parse --show-toplevel 2>/dev/null)
_SKILL_DIR=""        # this skill's own dir (for implement.md, fixtures/)
_PW_SKILL_DIR=""     # CJ_personal-workflow's dir

# Level 1: workbench repo
if [ -n "$_REPO_ROOT" ] && [ -f "$_REPO_ROOT/skills/CJ_implement-from-spec/implement.md" ]; then
  _SKILL_DIR="$_REPO_ROOT/skills/CJ_implement-from-spec"
fi
if [ -n "$_REPO_ROOT" ] && [ -f "$_REPO_ROOT/skills/CJ_personal-workflow/personal-artifact-manifests.json" ]; then
  _PW_SKILL_DIR="$_REPO_ROOT/skills/CJ_personal-workflow"
fi

# Level 2: deployed location
if [ -z "$_SKILL_DIR" ] && [ -f "$HOME/.claude/skills/CJ_implement-from-spec/implement.md" ]; then
  _SKILL_DIR="$HOME/.claude/skills/CJ_implement-from-spec"
fi
if [ -z "$_PW_SKILL_DIR" ] && [ -f "$HOME/.claude/skills/CJ_personal-workflow/personal-artifact-manifests.json" ]; then
  _PW_SKILL_DIR="$HOME/.claude/skills/CJ_personal-workflow"
fi

if [ -z "$_SKILL_DIR" ] || [ -z "$_PW_SKILL_DIR" ]; then
  echo "ERROR: Could not find skill assets."
  echo "Need: CJ_implement-from-spec AND CJ_personal-workflow assets."
  echo "Run: ./scripts/skills-deploy install (workbench) or check repo structure."
  echo "NOT_FOUND"
else
  echo "SKILL_DIR: $_SKILL_DIR"
  echo "PW_SKILL_DIR: $_PW_SKILL_DIR"
fi

If NOT_FOUND: tell the user "Error: CJ_implement-from-spec or CJ_personal-workflow skill assets not found. Run skills-deploy install or check the repo structure." and stop.

Overview

This skill takes a work-item directory of any type (user-story, defect, task, feature) and writes the code that its input artifacts describe:

• Per-type input artifacts (resolved from _TRACKER.md frontmatter type: field):

  Type	Input artifacts	Plan source
  user-story	SPEC.md + DESIGN.md (this story's + parent feature's)	SPEC's Components Affected + Data Flow
  defect	RCA.md + test-plan.md + TRACKER.md	RCA's Affected Components + Fix Description; test-plan rows define post-fix behavior
  task	TRACKER.md + test-plan.md	Tracker's Acceptance Criteria + Todos; test-plan rows define expected behavior
  feature	AUQ to pick a child user-story; delegate to that user-story's path	(delegated)

• Read context: input artifacts (per type) plus the parent feature's DESIGN.md if applicable.

• Plan the implementation: parse the input artifact's Components Affected (or equivalent) and the desired behavior. Detect input gaps ({placeholder} values, missing sections) and AUQ before proceeding.

• Sensitive-surface AUQ: if the input artifact names skills-catalog.json, any personal-artifact-manifests.json / company-artifact-manifests.json, or any validator script, AskUserQuestion before committing those writes.

• Default mode is propose-and-confirm: the skill writes a diff preview to chat and AskUserQuestion to approve / modify / cancel. --auto skips the preview for trivial changes (heuristic: ≤2 files touched AND no sensitive-surface change). --auto is per-invocation only in v1.

• Write code: Read/Edit/Write tools per the input artifact's architecture decisions.

• Update tracker: category-grouped journal entries ([impl-decision], [impl-finding], [impl], [impl-auto]); transition Phase 2 implementer-owned gates on green per the work-item's tracker template. For user-stories: Todos section reflects remaining work + Files section updated with changed files (the QA-owned gates Acceptance criteria verified met / Smoke tests pass are owned by /CJ_qa-work-item). For defects/tasks: per-template implementer-owned Phase 2 gates.

• Boundary check at start (Premise 1.3): runs /CJ_personal-workflow check on the work-item dir; refuses if Phase 1 isn't fully green.

• Boundary check at end (Premise 1.3): runs /CJ_personal-workflow check after writes; surfaces violations via AskUserQuestion.

• Idempotency (Premise 1.1): re-running on a work-item whose Phase 2 implementer-owned gates are already marked AND has a [impl-pass] journal entry is a NO-OP.

For the full step-by-step logic, see implement.md.

Usage

/CJ_implement-from-spec <work-item-dir>
/CJ_implement-from-spec <work-item-dir> --auto

The skill accepts any work-item type. The type: field in the work-item's _TRACKER.md frontmatter dispatches to the per-type input artifacts (see Overview table). If a feature directory is provided, the skill lists its child user-stories and AskUserQuestion which one to implement (preserves the v1.10.0 path).

--auto skips the propose-and-confirm preview when the change is trivial. The skill ignores --auto and falls through to propose-mode if the change touches a sensitive surface OR more than 2 files; the safety override is non-negotiable.

Examples (per type):

# user-story (today's path; unchanged)
/CJ_implement-from-spec work-items/features/CJ_personal-workflow/F000010_pipeline_skills/S000018_implement_from_spec
/CJ_implement-from-spec work-items/features/CJ_personal-workflow/F000010_pipeline_skills/S000018_implement_from_spec --auto

# defect (new path; reads RCA + test-plan)
/CJ_implement-from-spec work-items/defects/CJ_personal-workflow/D000016_test_deploy_stale_templates

# task (new path; reads TRACKER + test-plan)
/CJ_implement-from-spec work-items/tasks/CJ_personal-workflow/T000005_some_task

# feature (delegates to a child user-story via AUQ)
/CJ_implement-from-spec work-items/features/CJ_personal-workflow/F000012_pipeline_parity

Routing

Read implement.md and follow its instructions. The full implementation logic lives there: input validation, boundary check at start, idempotency, read context, SPEC gap check, plan, sensitive-surface AUQ, propose-and-confirm preview, write code, tracker updates, boundary check at end.

Error Handling