sec-filing-evidence-extractor

SEC Filing Evidence Extractor

Safety Notice

This listing is imported from skills.sh public index metadata. Review upstream SKILL.md and repository scripts before running.

Copy this and send it to your AI assistant to learn

Install skill "sec-filing-evidence-extractor" with this command: npx skills add toolanzyhhh1234/sec-shenanigans-agent-skills/toolanzyhhh1234-sec-shenanigans-agent-skills-sec-filing-evidence-extractor

SEC Filing Evidence Extractor

Extract section-anchored facts before interpretation.

Workflow

  • Identify filing set and periods.

  • Prefer latest annual plus at least 8 quarters.

  • Extract evidence rows with exact filing anchors and numbers.

  • Validate schema against bundle contract.

  • Write artifact evidence-table.md .

Required Output Artifact

Write evidence-table.md with required columns defined in ../sec-shenanigans-orchestrator/references/bundle-contract.md .

Quality Rules

  • Keep one factual disclosure per row.

  • Separate extraction confidence from manipulation confidence.

  • Preserve sign and units in numbers .

  • Mark ambiguous extraction with confidence=low .

Pre-Stage Input Helper

  • Local Priority: Always begin by scanning the workspace for a directory matching the /<ticker> (e.g., /PTC , /TSLA ). The user will often supply local 10-K , 10-Q , or .xlsx files in these nested directories. Treat these local files as your primary source of truth.

  • External Fallback: If local information is insufficient or missing, you are expected to proactively crawl the SEC website (EDGAR) or use provided fetching scripts to locate the necessary filings.

If raw statement tables are missing locally, orchestrator may run:

python skills/sec-shenanigans-orchestrator/scripts/fetch_financial_sheets.py <TICKER>

Use generated markdown only as a data helper; keep evidence anchored to SEC filing sections.

Use references/sec-filing-hotspots.md for section priorities.

Source Transparency

This detail page is rendered from real SKILL.md content. Trust labels are metadata-based hints, not a safety guarantee.

Related Skills

Related by shared tags or category signals.

Automation

sec-shenanigans-memo-writer

No summary provided by upstream source.

Repository SourceNeeds Review
Automation

sec-shenanigans-classifier

No summary provided by upstream source.

Repository SourceNeeds Review
Automation

sec-shenanigans-orchestrator

No summary provided by upstream source.

Repository SourceNeeds Review
Automation

sec-red-flag-ratio-checks

No summary provided by upstream source.

Repository SourceNeeds Review