code-add-observability

Structured logging at integration boundaries is added by instrumenting outbound HTTP, IPC, spawn/exec, and queue handoffs with leveled logs and typed error reporting. The Skill discovers your existing logger and injects recipe-driven event instrumentation directly into boundary code.

Preventing silent failures from empty catch blocks involves replacing swallow patterns with logged equivalents while preserving original semantics. The Skill targets empty catches and `|| true`-style behavior, adding structured error evidence so failures leave a debuggable trail.

Instrumenting webhook routes and file writes without new dependencies is possible because the Skill conforms to your project's existing logger. If no logger is found, it proposes a minimal zero-dependency fallback without installing anything first.

Sweeping a current branch diff for system boundaries is supported by the Skill. It enumerates inbound and outbound HTTP, IPC, spawn/exec, env injection, file writes, queues, hooks, and MCP tool handoffs, then applies recipe-driven instrumentation across all detected surfaces.

Greppable event names are needed so downstream debugging processes can quickly locate the evidence trail. The Skill documents where logs land and verifies at runtime that emitted event names are searchable, enabling a later reactive skill to find the trail reliably.

Observability instrumentation explicitly supports MCP tool handoffs and queue operations. The Skill targets these specific integration boundaries alongside HTTP, IPC, and file writes, adding structured leveled logs and typed error reporting to each surface.