Install
$ agentstack add skill-shalintripathi-organic-os-hoo-import-audit ✓ scanned · ✓ verified — works with Claude Code, Cursor, and more.
Security review
✓ PassedNo issues found. Passed automated security review. · v0.1.0 How review works →
- ✓ Prompt-injection patterns
- ✓ Secret / credential exfiltration
- ✓ Dangerous shell & filesystem operations
- ✓ Untrusted network calls
- ✓ Known-malicious package signatures
What it can access
- ✓ Network access No
- ✓ Filesystem access No
- ✓ Shell / process execution No
- ✓ Environment & secrets No
- ✓ Dynamic code execution No
From automated source analysis of v0.1.0. “Used” means the capability is present in the source — more access means more to trust, not that it’s unsafe.
About
Import an external audit (claude-seo shape)
Turns a point-in-time audit report into organic-os proposals without re-deriving the audit. The import trusts but attributes: findings keep their original evidence labels ([Measured]/[Inference]/[Unverified]), the original finding text rides in each proposal verbatim, and anything our own audit dimensions would dispute is flagged for the human, never silently rewritten.
- Input: a file path from the user (a FULL-AUDIT-REPORT.md, an
ACTION-PLAN.md, or similar markdown) or report text pasted in chat. Pasted text: write it to a temp file first. Read site-profile.yaml for the approval channel and the site URL.
- Parse and rank with the import module - stdlib only, defensive by
design (the format is theirs and may vary between versions; whatever matches no item pattern is preserved in parsed["unparsed"], never dropped): PYTHONPATH="$CLAUDE_PLUGIN_ROOT/lib" python3 -c snippet importing hoo.audit_import - parsed = parse_report(text), then props = to_proposals(parsed, max_items=5).
- Create each proposal via `create_item(kind=p["kind"], slug=p["slug"],
title=p["title"], body=p["body"], target=, source="import-audit"). createitem is the ONLY birth path - items are born proposed with an empty approvals list and are NEVER auto-approved; approval is the human's act, on the normal gate. Each body already carries the verbatim finding (quoted), its evidence label, the attribution line, and the dimension-mapping note: overlap names the onsite-audit dimension (skills/onsite-audit steps 3-4) via the DIMENSIONHINTS map in plugin/lib/hoo/audit_import.py`; an llms.txt finding gets the deliberate-skip note (our position in plugin/docs/evidence.md); no overlap is marked external-only and imported as-is on the source audit's evidence.
- Present a summary table in-session:
- imported items: title, source severity, dimension mapping
(dimension name | deliberate-skip | external-only)
- skipped items: title + reason (ranked below the max_items cut;
re-run with a higher max_items to import them)
- unparsed excerpt count, with a pointer that the excerpts were
preserved, not dropped
- Rebuild the queue and route the approval notification through the
configured channel exactly like any propose run (skills/onsite-propose step 4, including the isnotified/marknotified discipline so re-runs never re-notify).
- Append ONE signal recording the import: source audit date
(parsed["source_meta"]["audit_date"]), items parsed, proposals created, skipped count, unparsed excerpt count.
Pattern credit: claude-seo (https://github.com/AgriciDaniel/claude-seo), credited in README.md since v0.1.0. They audit, we operate.
Source & license
This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.
- Author: shalintripathi
- Source: shalintripathi/organic-os
- License: MIT
- Homepage: https://organicos.shivaatripathi.com
Install and usage instructions live in the source repository linked above.
Reviews
No reviews yet — be the first.
Write a review
Versions
- v0.1.0 Imported from the upstream source.