Fix gpclean: unrecognized codex model and silent no-safe-output stalls - #58725
Merged
Conversation
Co-authored-by: pelikhan <4175913+pelikhan@users.noreply.github.com>
Copilot
AI
changed the title
[WIP] Fix configuration issue for GPL Dependency Cleaner outputs
Fix gpclean: unrecognized codex model and silent no-safe-output stalls
Sep 5, 2026
pelikhan
marked this pull request as ready for review
September 5, 2026 04:56
Copilot stopped reviewing on behalf of
pelikhan due to an error
September 5, 2026 04:57
Contributor
There was a problem hiding this comment.
Note
Copilot could not run the full agentic suite for this review because it was automatically requested on a bot-authored pull request. Request a review from Copilot under Reviewers to retry with the full agentic suite. Improved support for bot-authored pull requests is coming soon.
Pull request overview
Updates the gpclean workflow skill configuration to use a newer Codex model identifier and improves agent-run completion/IO guidance so runs always end with a “safe output” (issue or noop) and avoid incorrect MCP file-reading attempts.
Changes:
- Switch model references from
openai/gpt-5.3-codextoopenai/gpt-5-codexin the prompt and locked workflow. - Add explicit guidance to read local SBOM/state files from disk (not via MCP resources).
- Require explicit termination via
noopwhen no GPL dependency is found, and expand “always finish with issue or noop” rules.
Show a summary per file
| File | Description |
|---|---|
| .github/workflows/gpclean.md | Updates model ID and tightens operational instructions around safe outputs and local-file access. |
| .github/workflows/gpclean.lock.yml | Regenerates lock metadata and updates all pinned model env vars to the new model ID. |
Review details
💡 Add a code-review agent skill for context-aware, tailored reviews. Learn more in the docs.
- Files reviewed: 2/2 changed files
- Comments generated: 2
- Review effort level: Lite
Comment on lines
+450
to
+453
| - If the SBOM file `/tmp/gh-aw/agent/sbom.json` is missing or corrupted, call `noop` explaining the error and stop (this should not happen as it's pre-downloaded in frontmatter) | ||
| - If `go mod graph` fails, call `noop` explaining the error and stop | ||
| - If license detection fails for a module, document it in the issue and recommend manual review | ||
| - If no direct dependencies exist, exit successfully | ||
| - If no direct dependencies exist, call `noop` explaining that there was nothing to check and stop |
|
|
||
| **IMPORTANT**: The SBOM has been pre-downloaded to `/tmp/gh-aw/agent/sbom.json` in the frontmatter setup step. **Use this file directly** - do NOT try to download it again using curl or gh api (you do not have a GitHub token in the agent environment). | ||
|
|
||
| **IMPORTANT**: These are plain local files on disk, not MCP resources. Read them with the bash tool (e.g. `cat /tmp/gh-aw/agent/sbom.json` or `cat /tmp/gh-aw/cache-memory/gpclean/state.json`) or a text-reading tool. Do NOT attempt to read them via `read_mcp_resource` or any MCP server (`github`, `local_cache`, etc.) - these files are not exposed as MCP resources and such calls will always fail. |
Contributor
|
🎉 This pull request is included in a new release. Release: |
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
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.
The GPL Dependency Cleaner (gpclean) workflow run produced no safe outputs: the configured codex model wasn't recognized by the installed CLI, and the agent got stuck trying to read local files as MCP resources before giving up without ever calling a safe-output tool.
Model fix
modelin.github/workflows/gpclean.mdfromopenai/gpt-5.3-codex(unrecognized by the installed Codex CLI, which fell back to generic model metadata) toopenai/gpt-5-codex, a stable, recognized Codex alias used elsewhere in the repo.Prompt robustness
/tmp/gh-aw/agent/sbom.json) and cache-memory state files are plain local files to be read withcat/bash — not viaread_mcp_resourceagainst thegithub/local_cacheMCP servers. The failed run repeatedly attempted the latter, which always errors with "unknown MCP server" / "Resource not found".go mod graphfailure, no dependencies, no GPL found) to explicitly callnoopwith a short explanation instead of vague "report the error and exit" wording.noop— never end silently.Regenerated artifacts
.github/workflows/gpclean.lock.ymlto reflect the model change and prompt updates.