How do I install the Sci Asset Manager skill?

Run npx skillmds@latest add godzhiwzz-create/sci-asset-manager in your terminal (requires Node.js), paste this page's agent-chat prompt into Claude, Cursor, or any MCP-connected agent, or download the SKILL.md file and copy it into your agent's skills directory.

Is Sci Asset Manager safe to use?

This skill has not completed SkillMD's automated safety review yet. SkillMD never runs a skill's scripts for you; review the SKILL.md before installing.

Which AI agents work with Sci Asset Manager?

This skill is tagged as working with Claude Code, Claude.ai, OpenAI Codex. SKILL.md is an open format, so most agents that read a skills directory can load it too.

Is Sci Asset Manager free to use?

Yes. Installing skills from SkillMD is free, and the skill stays under its author's original license.

Who published Sci Asset Manager?

godzhiwzz-create (@godzhiwzz-create) published this skill. Their other Agent Skills are listed on their SkillMD profile.

Sci Asset Manager by godzhiwzz-create

Name: Sci Asset Manager
Author: godzhiwzz-create

Sci Asset Manager

Review and maintain research project assets without losing evidence metadata, including cleanup candidates, cold archives, migration manifests, submission keep lists, and deletion risk. Use when deciding what to keep, move, archive, deduplicate, or delete in a research workspace. Default to bounded review; create durable manifests only when needed. Never delete or move protected assets without explicit user authorization.

File contents

SCI Asset Manager

Preserve evidence while reducing active-workspace clutter. Follow project-local rules and the provenance safeguards in sci-research-manager.

Read ownership before storage

Start from handoff, experiment index/registry, relevant cards, and claim map. Use known run paths before scanning large artifact directories. Never open large binary weights merely to classify them.

Always preserve

experiment/family card;
frozen config and code commit;
result summary and authoritative raw result;
run command/path;
archive/migration note;
claim-evidence relation;
original timestamp/size and optional hash for path-sensitive moves.

Potential cleanup candidates include reproducible caches, duplicate logs, temporary outputs, superseded generated files, and large intermediate checkpoints whose evidence metadata is preserved.

Review before action

Use the bundled templates for archive, cold-storage, submission-keep, and deletion reviews. Record path, owner, related ID/claim, reason, evidence risk, replacement, recommendation, and rollback.

Before an approved move, snapshot provenance with sci-research-manager/scripts/provenance_guard.py; prefer same-volume moves; update old-to-new mappings and links; then verify content, size, timestamp, symlink target, and optional hash.

Keep deletion separate

Never delete files unless the user explicitly authorizes the exact scope. First produce delete_review.md. Treat unknown ownership as needs_verification. A cleanup review is not deletion permission.

Review assets when requested or when a concrete storage, provenance, or outgoing-package risk requires it. A probe or phase change alone does not trigger an extra audit, manifest, archive, or deletion workflow. A narrow review may return findings in the conversation; create a durable manifest for an authorized migration/deletion or when the project requires one.

---name: sci-asset-managerdescription: Review and maintain research project assets without losing evidence metadata, including cleanup candidates, cold archives, migration manifests, submission keep lists, and deletion risk. Use when deciding what to keep, move, archive, deduplicate, or delete in a research workspace. Default to bounded review; create durable manifests only when needed. Never delete or move protected assets without explicit user authorization.---# SCI Asset ManagerPreserve evidence while reducing active-workspace clutter. Follow project-local rules and the provenance safeguards in `sci-research-manager`.## Read ownership before storageStart from handoff, experiment index/registry, relevant cards, and claim map. Use known run paths before scanning large artifact directories. Never open large binary weights merely to classify them.## Always preserve- experiment/family card;- frozen config and code commit;- result summary and authoritative raw result;- run command/path;- archive/migration note;- claim-evidence relation;- original timestamp/size and optional hash for path-sensitive moves.Potential cleanup candidates include reproducible caches, duplicate logs, temporary outputs, superseded generated files, and large intermediate checkpoints whose evidence metadata is preserved.## Review before actionUse the bundled templates for archive, cold-storage, submission-keep, and deletion reviews. Record path, owner, related ID/claim, reason, evidence risk, replacement, recommendation, and rollback.Before an approved move, snapshot provenance with `sci-research-manager/scripts/provenance_guard.py`; prefer same-volume moves; update old-to-new mappings and links; then verify content, size, timestamp, symlink target, and optional hash.## Keep deletion separateNever delete files unless the user explicitly authorizes the exact scope. First produce `delete_review.md`. Treat unknown ownership as `needs_verification`. A cleanup review is not deletion permission.Review assets when requested or when a concrete storage, provenance, or outgoing-package risk requires it. A probe or phase change alone does not trigger an extra audit, manifest, archive, or deletion workflow. A narrow review may return findings in the conversation; create a durable manifest for an authorized migration/deletion or when the project requires one.

Frequently asked questions