Instruction file imported from oldskultxo/aictx (.github/instructions/aictx.instructions.md). Copyright stays with the author.

AICTX Copilot path-specific lifecycle

Apply these instructions to repository work. Prefer AICTX MCP tools when available; fall back to AICTX CLI commands otherwise. They intentionally duplicate the minimal AICTX lifecycle because Copilot support varies by surface.

• MCP-first startup: if AICTX MCP tools are already visible, use those tools for resume/finalize. If they are not visible but .mcp.json or .vscode/mcp.json exists, first use the runner tool-discovery mechanism when available (for example search for aictx resume finalize lifecycle) so lazy-loaded MCP namespaces can attach, then have the runner attach/start the configured stdio MCP server before the first AICTX command of each new session. If MCP tools still are not attached after discovery/attachment, state that MCP config exists but tools are unavailable in this runner and use the CLI fallback.
• Start non-trivial repository tasks with: aictx resume --repo . --task "<task goal>" --agent-id copilot --adapter-id copilot-vscode --json.
• Use the resume capsule before broad repo scanning.
• Finalize non-trivial repository tasks with: aictx finalize --repo . --status success|failure --summary "<what happened>" --agent-id copilot --adapter-id copilot-vscode --json.
• Append the AICTX final summary from finalize when available. Render every provided summary section, including details, continuity_view_file, and continuity_view_online; preserve Continuity View file links and Mermaid online view links; do not replace URLs with placeholders and do not manually reconstruct or retype pako URLs.
• If command execution is unavailable, state that the AICTX lifecycle could not be executed.
• Do not inspect .aictx/ directly during normal startup.