From dstoic
Initialize OpenSpec in a project. Use when: user says 'init openspec', 'setup openspec', or asks to add OpenSpec to a project.
npx claudepluginhub digital-stoic-org/agent-skills --plugin dstoicThis skill uses the workspace's default tool permissions.
Initialize OpenSpec in a project. Detects framework and creates structure.
Creates isolated Git worktrees for feature branches with prioritized directory selection, gitignore safety checks, auto project setup for Node/Python/Rust/Go, and baseline verification.
Executes implementation plans in current session by dispatching fresh subagents per independent task, with two-stage reviews: spec compliance then code quality.
Dispatches parallel agents to independently tackle 2+ tasks like separate test failures or subsystems without shared state or dependencies.
Initialize OpenSpec in a project. Detects framework and creates structure.
Analyze project to identify framework and structure.
Steps:
package.json → Node.js (check for next, react, vue, svelte)pyproject.toml or requirements.txt → PythonCargo.toml → Rustgo.mod → GoOutput format:
framework: [detected]
language: [detected]
package_manager: [detected]
entry_points: [list]
Create OpenSpec directory structure.
Input: $ARGUMENTS = project description (optional)
Steps:
detect if not already doneopenspec/
├── project.md # From template + detected context
└── AGENTS.md # Copy from openspec CLI or template
project.md with:
Exploration Strategy template:
context_sources:
primary:
- openspec/project.md
- openspec/changes/{id}/proposal.md
- openspec/changes/{id}/tasks.md
- openspec/changes/{id}/test.md
- openspec/changes/{id}/specs/*.md
artifacts:
- openspec/changes/{id}/test-logs/*.md
must_read:
- CLAUDE.md
tools:
codebase: [Glob, Grep, Read, Task:Explore]
# mcp_serena: true # Uncomment for large C#/.NET
# mcp_sourcegraph: true # Uncomment for monorepos
Execution Philosophy template:
mode: garage # garage | scale | maintenance
modes:
garage:
description: Fast iteration, validate ideas quickly
principles:
- Ship fast, iterate faster (speed > polish)
- Working > perfect (80% solution now beats 100% later)
- Validate with real usage (dogfood immediately)
- Defer optimization until patterns stabilize
- Fail fast, learn faster
accept: [tech_debt, rough_edges, manual_steps, minimal_docs]
reject: [broken_core_flows, data_loss, security_holes]
anti_patterns:
- Over-engineering for hypothetical futures
- Gold-plating before validation
- Premature abstraction
- Analysis paralysis
scale:
description: Production hardening, team growth
principles:
- Stability over speed (measure twice, cut once)
- Document decisions for team alignment
- Abstractions justified by repeated patterns
- Test coverage for critical paths
accept: [slower_iteration, more_review, documentation_overhead]
reject: [untested_deployments, undocumented_apis, tribal_knowledge]
anti_patterns:
- Cowboy coding without review
- Skipping tests for speed
- Undocumented architectural decisions
maintenance:
description: Stability preservation, minimal changes
principles:
- If it works, don't touch it
- Minimal blast radius for changes
- Comprehensive regression testing
- Document before modifying
accept: [conservative_changes, extensive_testing, slow_rollouts]
reject: [unnecessary_refactors, feature_creep, risky_upgrades]
anti_patterns:
- Refactoring for aesthetics
- Upgrading dependencies without cause
- Adding features during maintenance windows
Ask user: "Which execution mode fits your current phase?" (default: garage)
Requires: User confirmation before writing files