Investigate safeinputs MCP server failure in dev.md workflow

[Copilot]

The dev.md workflow run failed during log parsing because Claude attempted to connect to a safeinputs MCP server that wasn't configured in the workflow's MCP server list.

Analysis

Workflow Context:

• Run: https://github.com/githubnext/gh-aw/actions/runs/20039762441
• Branch: main (poem creation workflow)
• Failure: "Parse agent logs for step summary" step

Root Cause:
Claude's initialization includes safeinputs in its MCP server list, but the compiled workflow (.github/workflows/dev.lock.yml) only configures github and safeoutputs servers. This causes the MCP server to fail during launch:

{
  "name": "safeinputs",
  "status": "failed"
}

The log parser detects this failure and fails the step with:

##[error]MCP server(s) failed to launch: safeinputs

Key Finding:
The agent successfully completed its task (created a poem), but the workflow failed during post-execution log parsing due to MCP server failure detection. This suggests safeinputs is being added to Claude's MCP server list even when not needed by the workflow configuration.

Notes

This issue is unrelated to the rest-api-assign-agent PR changes. The failed workflow run was from the main branch's dev.md (poem creation), not the PR branch's dev.md (issue assignment with assign-to-agent safe output).

---

✨ Let Copilot coding agent set things up for you — coding agent works faster and does higher quality work when set up for your repo.