[FEATURE] Link production incidents (Sentry/Datadog) back to the originating Claude Code session/commit

Summary

Claude Code's involvement in a codebase currently stops at code-gen/commit time — there's no mechanism linking a later production incident (an error in Sentry, Datadog, or similar) back to the Claude Code session or commit that introduced the underlying change. Please add an integration that closes this loop.

Current behavior

Once a commit lands, Claude Code has no ongoing relationship to that change. If it later causes a production incident, there's no path from "here's the error/alert" back to "here's the session and commit that produced the code responsible" — that correlation has to be reconstructed entirely by hand (via git blame, manual log correlation, etc.), even though Claude Code already has the session history that shows exactly what was done and why.

Motivation

Independent commentary on trusting AI-generated code in production (controltheory.com, "How to Trust AI-Generated Code in Production: A 2026 Guide") specifically identifies this as an unresolved gap across the AI-coding-agent market: tools stop at code generation and don't close the loop with real production/runtime feedback. Given Claude Code already tracks session-to-commit provenance internally (which session touched which files, when), the missing piece is specifically the runtime-incident side of that link, not the whole mechanism.

Proposed behavior

1. Optional integration with common observability tools (Sentry, Datadog, etc.) via their APIs/webhooks.
2. When an incident/error references a file or line, look up which Claude Code session(s) last touched that code and surface the link — "this error traces to changes from session X, commit Y."
3. Surface this either passively (available on request, "what session produced this code") or actively (a notification when a tracked incident matches recently-touched code) — the passive, on-request version alone would already close most of the gap.

Why it matters

• Extends Claude Code's existing session-provenance tracking to cover the part of the lifecycle that actually matters most for trust — whether a change caused a real problem, not just whether it compiled and passed tests at commit time.
• Turns "figure out which change broke prod" from manual archaeology into a direct lookup, for exactly the kind of code Claude Code itself helped write.

Related issues (adjacent, not duplicates)

None found describing this specific integration.

Notes

• Reviewed for duplicates before filing (multiple search passes: "link production incident error Sentry Datadog back to commit session", "observability production bug trace to originating change") — nothing found as of 2026-07-19.
• Sourced from market research: controltheory.com, "How to Trust AI-Generated Code in Production: A 2026 Guide," describing this as a general, unresolved gap across AI coding agents rather than something specific to a personal incident.