Skip to content

openJiuwen-ai/sciencediscovery / skills/antibody-design

antibody-design

Unverified

Prepare, launch, monitor, and summarize the real RFdiffusion to ProteinMPNN to Protenix antibody pipeline on a local or remote ScienceDiscovery Runner with sandboxed Ascend NPUs.

AGENT_BOOTSTRAP.md

View on GitHub
# Agent bootstrap: sandboxed Protenix antibody run

1. Choose one authorized local or remote NPU Runner and use its `runner_id` for
   every environment, Shell, and workspace-sync call.
2. Require the target PDB, framework PDB, and explicit chain-labelled hotspot
   list before preparing a run. If any is missing, ask the user and stop; do not
   infer hotspots or delegate the decision to a SubAgent.
3. Select and probe a managed scientific environment on the same Runner.
4. Create the workspace-relative config, then launch the frozen Skill
   entrypoint once with `--prepare-only` and `background=true`. The first run
   clones the pinned official MindScience and RFdiffusion `sharker` sources and
   applies the bundled RFdiffusion MindSpore Tensor-to-PDB compatibility patch
   before downloading the three official RFdiffusion, ProteinMPNN, and
   Protenix checkpoints; existing checkouts are
   verified/re-pinned and verified files are reused in the same Session. The
   sandbox network allowlist must contain `gitcode.com`, `gitee.com`,
   `tools.mindspore.cn`, and
   `af3-dev.tos-cn-beijing.volces.com`. Keep the returned Execution ID and
   monitor it; never duplicate a queued, running, or unknown preparation.
5. Validate in the foreground, then launch the frozen Skill entrypoint once with
   `run_shell(scriptPath="$SCIENCEDISCOVERY_SKILLS_DIR/antibody-design/scripts/run_sandbox_pipeline.sh", arguments=["--config", "antibody_pipeline/config.json"], runner_id=..., environment_id=..., background=true)`.
6. Retain the returned Execution ID. Wait and monitor with
   `execution_status` / `execution_logs`; never use `nohup`, shell sleep loops,
   file polling, `run_npu_job`, or automatic replay.
7. Selected host NPU cards are renumbered inside the sandbox. Use logical
   `npus="0"` for one selected card or `"0,1,..."` for multiple selected cards.
8. Remote outputs must be pulled with `sync_remote_workspace` before they can be
   declared as local Project artifacts.
writing-skillsobra/superpowersUse when creating new skills, editing existing skills, or verifying skills work before deploymentStars: 322kUpdated Claude CodeCodexVerifiedInstall
xlsxanthropics/skillsUse this skill any time a spreadsheet file is the primary input or output. This means any task where the user wants to: open, read, edit, or fix an existing .xlsx, .xlsm, .xltx, .csv, or .tsv file (e.g., adding columns, computing formulas, formatting, charting, cleaning messy data); create a new spreadsheet from scratch or from other data sources; or convert between tabular file formats. Trigger especially when the user references a spreadsheet file by name or path — even casually (like "the xlsx in my downloads") — and wants something done to it or produced from it. Also trigger for cleaning or restructuring messy tabular data files (malformed rows, misplaced headers, junk data) into proper spreadsheets. The deliverable must be a spreadsheet file. Do NOT trigger when the primary deliverable is a Word document, HTML report, standalone Python script, database pipeline, or Google Sheets API integration, even if tabular data is involved.Stars: 198kUpdated Claude CodeCodexVerifiedInstall
shipping-and-launchaddyosmani/agent-skillsPrepares production launches. Use when preparing to deploy to production, or when asking what needs to be in place before shipping. Use when you need a pre-launch checklist, when setting up monitoring, when planning a staged rollout, or when you need a rollback strategy.Stars: 119kUpdated Claude CodeCodexVerifiedInstall
childhoodthedotmack/claude-memBabysit a shipped release after merge, publication, or deployment. Verify published packages, production behavior, installed plugins, and late CI or user reports until the observation window is clean; use babysit for PR readiness before release.Stars: 118kUpdated Claude CodeCodexVerifiedInstall