Skip to main content

/build-plan

Source: ~/.claude/skills/build-plan/SKILL.md

name: build-plan version: 2.0 description: > Execute an approved plan using TaskList with builder/validator teams. Use after /plan-with-team creates an approved plan. Triggers: "execute the plan", "build this", "implement the plan", "run the plan", "/build-plan". argument-hint: "[path to plan file, or leave empty for latest in ~/system/specs/]" level: 3

/build-plan — Execute Approved Plan

Purpose

Executes a pre-approved plan file using parallel builder/validator TaskList agents. Run AFTER /plan-with-team — never execute without an approved plan.

Variables

Variable Type Description Default
plan_path path Path to plan file latest in ~/system/specs/
concurrency number Parallel builders 3
yolo flag Skip browser testing false

Workflow

Step 1: Load Plan

  • if $ARGUMENTS provided → use as plan_path
  • else → find latest plan: ls -t ~/system/specs/*.md | head -1
  • Read plan file, verify it has: [ ] task checkboxes + acceptance criteria

Step 2: Validate Prerequisites

  • Check: does plan have approval marker? (## APPROVED section or status: approved)
  • if not approved → STOP, tell user to run /plan-with-team first
  • Check: are required services running? (DB, relevant tools)

Step 3: Activate Build Mode

node ~/system/tools/build-mode.js start <project_dir> --concurrency <N>

Builder completion flow

Builders call node ~/system/tools/mc.js ready <id> NOT mc.js done. Only after Proveo/validator verification can tasks be marked done. Build-plan CANNOT report COMPLETE until all tasks pass Proveo verification.

Step 4: Execute Tasks (parallel where independent — with Context Isolation)

For each [ ] task in plan, dispatch builders using ISOLATION PATTERN:

  1. Load ~/system/prompts/SUBAGENT_ISOLATION.md
  2. Instruct builder: "Write FULL implementation notes to /tmp/{task_id}-build-deliverables.md"
  3. Request: "Return ONLY: summary (what was built) + deliverable path + memory_candidates"
  4. if task has no dependencies → spawn in parallel
  5. if task has dependencies → wait for dep completion, then spawn

Token savings: Each builder returns ~200 tokens instead of 50K-200K implementation logs.

Step 5: Validate Each Task (with Context Isolation)

After each task completes → spawn validator agent using ISOLATION PATTERN:

  1. Load builder summary + optionally Read builder deliverable if needed
  2. Dispatch validator (Proveo/Angie Jones) with isolation preamble:
    • "Write FULL test results to /tmp/{task_id}-validation-deliverables.md"
    • "Return ONLY: PASS/FAIL + summary + deliverable path"
  3. Runs tests relevant to task
  4. Checks acceptance criteria
  5. if FAIL → send back to builder with failure context (Read validation deliverable for details)
  6. if PASS → proceed

Token savings: Validator returns ~150 tokens instead of 30K-100K test logs.

Step 6: Final Check

  • All tasks [x]? → run full test suite
  • if PASS → node ~/system/tools/build-mode.js stop --status completed
  • if FAIL → node ~/system/tools/build-mode.js stop --status failed + report

Report Format

BUILD-PLAN EXECUTION REPORT
Plan: [filename]
Status: [COMPLETE|PARTIAL|FAILED]
Tasks: X/Y completed
Tests: [passing/total]
Duration: [time]
Issues: [any blockers or failures]

$ARGUMENTS

Back to top