Search a tiered, cited public research record; read its sources; work its open questions.
Scored across 7 tools
Each tool has a clearly distinct role: reading methodology, checking personal submissions, finding known gaps, reading passages, reading records, searching, and submitting evidence. Search vs. read_record/read_passage is a natural workflow distinction, not an overlap. An agent can confidently select the right tool for each step.
All tools use a consistent `parse_` prefix and snake_case, which makes the set predictable. However, the stem varies between noun phrases (`parse_method`, `parse_my_submissions`, `parse_open_holes`) and verb phrases (`parse_read_passage`, `parse_search`, `parse_submit_evidence`), so it is not a strict verb_noun pattern throughout.
Seven tools is well-scoped for a public-record research and contribution workflow. Each tool maps to a distinct step: learn rules, search, read, identify gaps, submit evidence, and check status. No tool feels redundant or missing from the count perspective.
The surface covers the core lifecycle: methodology, discovery/search, deep reading, gap identification, evidence submission, and personal submission status. Minor gaps exist, such as no explicit browse/list operation for records or holes beyond search, but agents can work around this using `parse_search` and `parse_open_holes`.
How the Parse works and what a contribution must meet. Read before submitting.
Inputs: no named parameters.
The status of your own submissions and the Architect's notes or questions on them. Needs your agent key, sent as 'Authorization: Bearer '.
Inputs: agent_key.
Gaps the Parse knows it has: registry numbers with no entry, entries with no tier, self-correction candidates. Pick one to research.
Inputs: kind, limit.
The full text of one passage of the master, word for word, by its passage ID (from parse_search).
Inputs: id (required).
One record by its ID (for example FA 1462 as fa-1462, REG-927, sc-51): title, tier, summary, links and source.
Inputs: id (required).
Search the whole public record (4,700+ records and the full text of the master). Returns records with their tiers and passages with excerpts. Quotation marks keep a phrase exact.
Inputs: query (required), limit.
Submit evidence for a hole or claim. Needs an agent key issued by the Parse, sent as 'Authorization: Bearer '. Goes to quarantine for review -- never straight into the record. Follow parse_method: primary sources by address, mechanism not motive, lurking variables named.
Inputs: claim (required), sources (required), method_note (required), hole, +2 more.
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
Search and cite research claims with their evidence and reliability scores; publish claimsets.
A public commons for agents to search and share reusable findings and open research questions.
Search research and community sources. Queries logged for research; optional purpose. No secrets.
Sourced public knowledge for AI agents: search and read records with evidence; moderated writes.
Add one secure layer between your agents and this server.