Skip to content

🤖 docs: bug-bash container runner #5932

Description

@ThomasK33

Docs for the bug-bash container runner. Part of #5714. It was planned as part of B1 PR 4b, which stopped with #5883.

Owner: the coordinator desk (workspace 3203e76efb), assigned to ThomasK33.
Trigger: with #5931 (the CI job), in the same change.

Scope:

  1. The "Bug Bash and Repro Tests" section of docs/AGENTS.md: how to run repros through the runner, what it refuses (Docker Desktop, rootless Docker, non-mock modes), and its exit codes (2 refused, 3 unknown cleanup state, 130/143 stopped).
  2. What stays out of scope: the 🤖 tests: pause model-driven bug-bash runs on the host #5815 host pause and the mock-only boundary stay, and crash recovery is 🤖 tests: recover bug-bash sandbox containers left by a launcher crash #5882.
  3. Regenerate builtInSkillContent.generated.ts after the docs edit.

Depends on #5930 for the final runner behavior.


Generated with xum • Model: anthropic:claude-opus-5-5 • Thinking: high

Activity

  1. self-assigned this
    on Oct 9, 2026
  2. ThomasK33 commented on Oct 9, 2026

    @ThomasK33
    MemberAuthor

    Owner: the coordinator desk (workspace 3203e76efb). Next step: Docs for the runner. Trigger: with #5931..

    The coordinator desk assigned this owner on 2026-10-09 after an ownership sweep of open issues. If this owner stops, the coordinator desk takes the issue back.


    Generated with xum • Model: anthropic:claude-opus-5-5 • Thinking: high

  3. ThomasK33 commented on Oct 11, 2026

    @ThomasK33
    MemberAuthor

    Done on main. The bug-bash rules in docs/AGENTS.md (the host pause, BUGBASH_BUDGET_USD, the sandbox exit codes, the sandbox checks, what the launcher refuses, and recovering leftovers) came with the #5714 stack (#6073, #6076, #6079, #6082, #6085). #6095 (merged as ec554b9acf) corrected the one wrong line: make bug-bash-sandbox-check launches three jobs, not one. The generated skill file mirrors the docs.


    Generated with xum • Model: anthropic:claude-opus-5-5 • Thinking: high

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

Labels

featureTriage: feature request, not a bug

Type

No type

Projects

No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions