get-financials — independently scanned and version-tracked by SaferSkills.
SaferSkills independently audited get-financials (Agent Skill) and scored it 100/100 (green). The audit ran 55 deterministic rules across Security, Supply Chain, Maintenance, Transparency, and Community; it found 0 high-severity and 0 lower-severity findings. The full rule-by-rule trace and per-finding evidence are below. Free, methodology-open.
Findings & checks · 0 flagged
Every scanned point with the score it earned and what moved between them.
First recorded scan — no prior version to compare against.
The primary manifest — the file an agent reads to learn what this artifact does.
Pull the latest annual accounts and extract headline numbers.
get-filingsget-filings then fetch the specific docWhat was Microsoft Ltd's revenue last year? Company is GB 01624297.
Pull the last 3 years of Samsung Electronics accounts (KR).1. If user only gave a name, run find-company first.
2. list_filings({ jurisdiction, company_id, category: "accounts", limit: 3 })
→ capture document_id of the most recent (or top N for multi-year)
3. get_document_metadata({ jurisdiction, document_id })
→ check size + available_formats before fetching
4. fetch_document({ jurisdiction, document_id })
- kind='embedded': bytes_base64 is the full doc; read directly
- kind='resource_link': oversized PDF. Use get_document_navigation to find
the relevant section, then re-call fetch_document(pages='N-M', format='pdf')
5. EXTRACT from the bytes (record the XBRL concept URI for each figure when present):
- Revenue / Turnover
- Operating profit, PBT, Net income
- Total assets, Equity, Cash
- Average employees, Dividends declared
- Going-concern note, qualified audit opinion flag
6. REPORT (tabular):
- Concept | Value | Currency | Year | document_id | page
- Cite document_id + page number on every figure.has_document: false or empty available_formats): surface the upstream's purchase link, do not fabricate numbers.format='png' per page; warn the user that figures need manual verification.fetch_document fails (rate-limit / 5xx / timeout), do NOT fill in numbers from training data. Tell the user what failed and offer retry or source_url_official.get_document_navigation; always quote from the actual fetched page bytes.~30 seconds. Free. No account. Every finding cites a rule and a line of evidence.