Draft, confirm, and synchronize research status, evidence, blockers, and review-ready GitHub summaries using the Research Memory skill, GitHub conventions, and file templates.
The English summary is being prepared. Showing the original description.
Review the instructions the agent will receive before you install.
Coordinate research memory updates through this skill, GitHub conventions, and repository templates.
Use this skill when the user asks to record, update, summarize, or synchronize:
Do not use this skill for ordinary code edits, local-only notes, or GitHub actions that do not involve research memory.
Core Rule
The agent MUST follow this order for every Research Memory update:
If the user asks to "just sync it" without seeing a draft, still produce a draft first and request confirmation.
No MCP
Research Memory is intentionally a skill plus conventions now. Do not call, configure, install, or recreate a project-memory MCP server or worker. Use normal file edits and GitHub issue/PR/comment workflows after confirmation.
Templates
Prefer the repository templates in docs/templates/research-memory/:
artifact-record.mdgithub-issue.mdgithub-pr.mdstatus-update.mdFill only fields supported by the user's confirmed facts. Leave unknown fields as TBD or ask a short follow-up question when the field is required for GitHub output.
Hard Limits
The agent MUST NOT use GitHub or any connected workflow to:
validatedWhen a user asks for one of these actions, draft the proposed wording or action and ask for explicit confirmation. Treat validated and public claims as high-risk labels requiring clear user authorization and supporting evidence.
GitHub Requirements
All GitHub-facing summaries created from Research Memory MUST include:
Use concise labels such as:
Artifact ID: EXP-2026-06-25-rerank-a17Evidence level: preliminaryIf either value is missing, ask for it before posting or preparing a PR.
GitHub PRs carry review. Use PR comments, PR descriptions, or review threads for critique, questions, and reviewer decisions. Do not treat a status page as the review channel.
Evidence Levels
Prefer one of these evidence levels unless the project defines a stricter vocabulary:
observation: seen in a run, log, issue, or collaborator reportpreliminary: supported by an artifact but not independently checkedreproduced: repeated with matching outcomevalidated: explicitly approved by the user as validatedNever infer validated from passing tests, a successful run, or a collaborator comment.
Workflow
Minimum Examples
Experiment completed:
Draft memory update:
Artifact ID: EXP-2026-06-25-rerank-a17
Evidence level: preliminary
Summary: Reranker A17 completed on the held-out query set. It improved top-3 recall from 0.71 to 0.76, with latency increasing from 84 ms to 109 ms.
Evidence: run run-4821, metrics artifact metrics/rerank-a17.json
Target: Local artifact record and GitHub PR summary
Please confirm before I update files or prepare GitHub output.
Blocker update:
Draft memory update:
Artifact ID: RUN-2026-06-25-cuda-repro
Evidence level: observation
Summary: CUDA reproduction is blocked because the available runner has driver 555.42, while the failing environment used 560.31. Reproduction should wait for a matching runner or a containerized driver matrix.
Evidence: issue #214, CI log cuda-repro-993
Target: Local artifact record only
Please confirm before I update the record.
Collaborator issue feedback:
Draft GitHub issue response:
Artifact ID: DOC-2026-06-25-eval-feedback
Evidence level: observation
Summary: Collaborator feedback notes that the current eval excludes multi-hop retrieval failures. Proposed follow-up is to add a multi-hop slice before treating the reranker result as validated.
Evidence: collaborator comment on issue #219
Target: GitHub issue comment and local artifact record
Please confirm before I update files or post the issue response.
Output After Synchronization
Report what changed and where:
Research Memory update synchronized.
- Artifact ID: <id>
- Evidence level: <level>
- Targets updated: <local file | GitHub issue | GitHub PR | PR comment | status update>
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.
Submit protein folding tasks to the AlphaFold3 C550 inference platform. Use when the user needs to run AlphaFold3 predictions with pre-prepared JSON input files on the AI4S Kubernetes cluster.
The English summary is being prepared. Showing the original description.
Build and sharpen a project's domain model. Use when discussing codebase terminology, writing or editing a CONTEXT.md, or recording or editing an ADR.
The English summary is being prepared. Showing the original description.
A relentless interview to sharpen a plan or design, which also creates docs (ADR's and glossary) as we go.
The English summary is being prepared. Showing the original description.
Grill the user relentlessly about a plan, decision, or idea. Use when the user wants to stress-test their thinking, or uses any 'grill' trigger phrases.
The English summary is being prepared. Showing the original description.
Parse PDFs, Office files, images, and HTML into Markdown/structured outputs with MinerU. Use when SciForge or Codex needs document parsing/OCR for scientific papers, supplementary files, PDFs, scanned documents, tables, formulas, or URL/local-file parsing through MinerU standard or Agent lightweight APIs.
The English summary is being prepared. Showing the original description.
Implement tasks from an OpenSpec change. Use when the user wants to start implementing, continue implementation, or work through tasks.
The English summary is being prepared. Showing the original description.