Tools provided
Tools the agent exposes — your AI client calls them automatically when it needs them.
- run_full — Orchestrated path: resolve the bug context, build the minimal reproduction, generate the failing test, locate the affected code path, and return the full reproduction plan.
- analyze — Run pure deterministic analysis over a provided bug context with no model call and return scored reproduction-readiness and risk findings.
- build_repro_plan — Build the full deterministic reproduction plan: minimal steps, a failing test, the affected code path, expected versus actual, a suggested fix and a confidence score.
- generate_failing_test — Generate only the failing test artifact — framework, filename and code. The evidence produced before any fix, and red by contract.
- repro_steps — Return only the minimal, ordered reproduction steps for the provided bug context.
- locate_affected_path — Locate the affected code path: the file and line, the error type, the triggering input, expected versus actual, and the suggested fix.
- score_repro_confidence — Score the reproduction confidence from 0 to 1, flag environment-dependent bugs, and return the underlying risk signals.
- list_capabilities — List the agent's static capabilities: the available tools and the analyzer's risk-signal thresholds. Useful for discovery.
- plan_inputs — Plan the inputs for a tool: returns the clarifying questions, the JSON schema and a ready-to-edit example.
- discover_intent — Understand your goal and co-design the exact input through clarifying questions before running.
- open_form — Opens the guided form to collect the ticket id, the free-text report and an optional component hint. The entry point for running the agent on your own data.
- run_form — Internal: invoked by the guided form when the user submits. Not for direct use.
