Add OrcaReplay under MCP Servers - #455
Open
xizhuomengcontin wants to merge 1 commit into
Open
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Adds OrcaReplay — one line, in the section this fits.
Disclosure: I work on it. Apache-2.0, free, no paid tier, no account.
What it is
A record/replay debugger for coding agents. It captures a run below the harness — a base-URL variable moved for that process, or TLS terminated for one named host — so the agent itself is unmodified and needs no SDK, exporter or plugin. The trace keeps the verbatim request and response bytes, which is what makes the next part possible: the run re-executes with the network off (
egress=blocked), served the recorded exchanges back.Why it is not the same thing as the tracing tools already listed
The OpenTelemetry path — which is what Langfuse, Phoenix and the rest consume — does not capture message content by default, and even with
OTEL_INSTRUMENTATION_GENAI_CAPTURE_MESSAGE_CONTENT=truewhat lands is span attributes: a normalised view of the call, not the bytes on the wire. You cannot serve a span back to an SDK as an HTTP response. It also assumes the agent can be instrumented, and the agents most people actually run — Claude Code, Codex CLI, goose — are binaries you invoke, not libraries you import.So this sits beside those tools rather than replacing them: they are the system of record for a fleet, this is a debugger you point at one run.
Verified, not asserted
Two harnesses have been driven end to end against the real thing, and each broke something that got fixed:
1 failed→ agent fixes it →1 passed, then strict offline replayreused=8/8 exact=8 divergences=0, three runs in a row.reused=1/1 egress=blocked.Every adapter has a fixture recording the exact environment variables it sets, so a harness that renames one turns a check red instead of producing an empty trace.
Details
npm i -g orcareplayorca mcp, protocol2025-06-18) and an Agent SkillHappy to reword, shorten, or move it to a different section — say which and I will push the change.