Federal Register Doc Fetch

Core Goal

• Search the Federal Register documents API with one bounded request plan.
• Retrieve official notices, proposed rules, final rules, or presidential documents relevant to one mission topic.
• Return machine-readable JSON records with publication dates, agencies, URLs, excerpts, docket IDs, and RIN metadata.
• Keep execution deterministic with retries, throttling, pagination caps, and payload validation.

Repository Policy

• This is the only Federal Register source skill in this repository.
• Do not create or invoke parallel Federal Register wrappers for the same documents endpoint.
• When eco-council or OpenClaw assigns a raw artifact path, write this skill's full JSON payload to that exact path with --output.

Required Environment

• Configure runtime by environment variables in references/env.md.
• Start from assets/config.example.env.
• Load env values before running commands:

set -a
source assets/config.example.env
set +a

Workflow

1. Validate effective configuration.

python3 scripts/federal_register_doc_fetch.py check-config --pretty

2. Dry-run the search plan before making remote calls.

python3 scripts/federal_register_doc_fetch.py fetch \
  --term "wildfire smoke EPA" \
  --start-date 2023-06-01 \
  --end-date 2023-06-10 \
  --agency environmental-protection-agency \
  --document-type NOTICE \
  --max-pages 2 \
  --max-records 20 \
  --dry-run \
  --pretty

3. Fetch one bounded search window and write the payload.

python3 scripts/federal_register_doc_fetch.py fetch \
  --term "wildfire smoke EPA" \
  --start-date 2023-06-01 \
  --end-date 2023-06-10 \
  --agency environmental-protection-agency \
  --document-type NOTICE \
  --output ./data/federal-register-docs.json \
  --pretty

4. Use task-specific structured filters when the mission already knows them.

python3 scripts/federal_register_doc_fetch.py fetch \
  --term "greenhouse gas" \
  --start-date 2024-03-01 \
  --end-date 2024-03-31 \
  --regulation-id-number 3235-AM87 \
  --document-type RULE \
  --document-type PRORULE \
  --max-records 50 \
  --pretty

Output Record Shape

Each item in records is one Federal Register document record, typically including:

• document_number, title, type
• publication_date, effective_on
• agencies, topics
• abstract, excerpts
• html_url, pdf_url, raw_text_url, comment_url
• docket_ids, regulation_id_numbers
• source_query_url, source_page_number

The full raw payload also keeps request metadata, page summaries, and validation output for downstream auditing.

Scope Boundaries

• This skill targets the Federal Register documents search endpoint only.
• This skill does not crawl linked HTML, PDF, or raw-text bodies.
• This skill does not infer legal meaning or summarize regulations.
• This skill does not need an API key.
• This skill is the canonical Federal Register fetch interface for this repository.

References

• references/env.md
• references/federal-register-api-notes.md
• references/federal-register-limitations.md
• references/openclaw-chaining-templates.md

Script

• scripts/federal_register_doc_fetch.py