| name | symptomsync-agentic-ai-change |
| description | Guide for changing SymptomSync's Python symptom-analysis service under agentic_ai/. Use when tasks touch the FastAPI service, LangGraph flow, agents, chains, settings, tests, monitoring, or deployment configuration for the agentic_ai workspace. |
SymptomSync Agentic AI Change
Use this skill when the task is primarily in agentic_ai/.
Workflow
- Read
agentic_ai/AGENTS.md.
- Classify the task:
- API surface
- graph or state logic
- agent behavior
- runtime settings
- observability or deployment behavior
- Trace the smallest set of files that share the contract before editing.
- Verify behavior from code rather than relying on README claims.
Important repo truths
- The checked-in service is a FastAPI app with
/health, /metrics, and /api/v1 routes.
- Settings live in
config/settings.py and use the SYMPTOMSYNC_ env prefix.
- Route or model changes usually require touching
mcp_server/models.py, mcp_server/routes.py, and tests together.
- Monitoring behavior is part of the contract, not optional polish.
Read references/service-map.md when you need the main service hotspots.
Validation
- Run
make lint in agentic_ai/.
- Run
make test in agentic_ai/.
- If only a subset is relevant, inspect the matching tests even if full execution is not possible.
Guardrails
- Do not advertise true MCP protocol behavior unless the current code actually implements it.
- Keep request and response shapes synchronized across models, routes, and tests.
- Be careful with env variable names; the settings object expects
SYMPTOMSYNC_ prefixes.
Related skills
- Use
$symptomsync-release-change when service changes also affect deployment or runtime config.
- Use
$symptomsync-doc-sync when service docs or operational docs need updating.
