Implement and land a fix for an explicitly identified marin-community/marin GitHub issue.
The English summary is being prepared. Showing the original description.
Review the instructions the agent will receive before you install.
Read AGENTS.md and the issue guidance in writing-style. Keep issue comments
terse: omit filler and repetition, use annotated links, and keep each prose
section to three or four sentences. Complete the workflow in order. If blocked,
comment on the issue with the last completed state and the blocker.
Use gh to fetch the issue. Read the codebase for all relevant source files.
Post one concise comment with Research and Proposed Fix sections.
Before writing code, run a duplicate-work preflight:
gh pr list --state open --search "<issue-id or keyword>").gh issue list --state open --search "<keyword>").In Research, state the cause in two or three sentences and link at most five
relevant code locations with short annotations. In Proposed Fix, name the
smallest fix and include the failing call chain or a snippet only when needed.
Implement on agent/{YYYYMMDD}-fix-{issue-id}.
Follow write-tests. Prefer a regression test that fails before
the fix, extend an existing file, run the narrow test and affected safe suite,
then run ./infra/pre-commit.py --all-files --fix. Do not override the configured
marker expression.
When all tests pass, upload your branch and open a PR following
.agents/skills/commit/SKILL.md exactly — use the plain-text format it
specifies (no markdown headers, bullet lists, or ## Summary sections;
violations are rejected). Attach a comment to the Github issue summarizing the fix.
After opening the PR, follow step 9 of .agents/skills/commit/SKILL.md
exactly. Its wait_for.py loop owns CI, review feedback, and lifecycle events.
Do not start a separate gh pr view or gh pr checks polling loop. Investigate
failures, push fixes, and re-arm the wait as that skill specifies.
Post a final issue comment with the landed PR or the last terminal state.
No reviews yet. Tell others what you thought.
No login needed. Share how it worked for you, ask a question, or note where it did not run.
No comments yet. Be the first to write one.
Run an explicitly requested Zephyr control/treatment benchmark on the same pre-normalized sample and compare Finelog stage metrics.
The English summary is being prepared. Showing the original description.
Register a named Hugging Face dataset for Marin by inspecting its schema and adding the appropriate experiments/datasets module.
The English summary is being prepared. Showing the original description.
Implement a fully specified request for a GCP IAM resource grant or Cloud Run IAP viewer in marin-iac.
The English summary is being prepared. Showing the original description.
Add or change a named Pallas/Mosaic kernel, including its reference implementation, correctness tests, wrapper, or requested tuning.
The English summary is being prepared. Showing the original description.
Archive explicitly named legacy experiment files by publishing a dated tag and canonical issue breadcrumbs.
The English summary is being prepared. Showing the original description.
Launch or continuously monitor a specified Zephyr pipeline on Iris only when asked to babysit it; restart only after explicit approval.
The English summary is being prepared. Showing the original description.