| name | ai-context-mcp |
| description | Use when working with ai-context MCP, .context scaffolding, PREVC workflow, plans, agent orchestration, semantic context maps, skills, docs, or context sync. |
ai-context MCP
Use ai-context for structured project context and PREVC workflows.
Repository setup
- Always pass
repoPath: D:\\projetos\\cifrashub on the first context call, usually context check.
- If
.context is missing and the task is non-trivial, initialize it before workflow setup.
- After initializing scaffolding, fill relevant scaffold files before relying on them.
PREVC workflow
- Use workflow init for non-trivial features/fixes.
- Scale guidance:
- QUICK: tiny/single-file changes, Execute + Verify only.
- SMALL: simple feature, Plan + Execute + Verify.
- MEDIUM: design decisions, Plan + Review + Execute + Verify.
- LARGE: security/compliance/systems, full Plan + Review + Execute + Verify + Complete.
- Respect gates unless user explicitly approves bypass/autonomous mode.
- Link plans when a plan exists and update phase/step status as work progresses.
Planning and agents
- Use
agent orchestrate or getSequence when multiple specialized roles would help.
- Use plan tools for decisions, phase status, step outputs, and markdown sync.
- Use skills tools to inspect or scaffold ai-context skills, not opencode skills.
Semantic context
- Use
buildSemantic for focused context around files or documentation/playbook generation.
- Use
getMap for stack, structure, architecture, symbols, public API, dependencies, or stats.
- Keep generated artifacts inside
.context/.
Boundaries
- ai-context is for workflow/context, not direct source edits.
- Use Serena or file edit tools for actual code changes.