What problem does it solve? When a source is about to be cited or relied on, gut feelings about trustworthiness are not defensible. This Skill applies the NATO Admiralty Code (STANAG 2511 / AJP-2.1, FM 2-22.3 Appendix B) to grade a source's reliability (A–F) separately from the credibility of the specific information it reports (1–6), producing a combined code like B2 with a one-sentence rationale. ## Core Features & Use Cases - Two-axis grading: Assigns a reliability letter from the originator's track record and a credibility digit from corroboration and logic, keeping the axes independent instead of collapsing them into one score. - Structured JSON output: Emits a record with source, reliability, credibility, combined code, rationale, and grading date for each graded source. - Deterministic companion tool: scripts/admiralty.py (Python 3, stdlib only) validates codes, prints the 6×6 matrix, aggregates grades across sources about one claim, and flags conflicts between well-graded sources. - Use Case: A draft cites a Reuters report of an unconfirmed funding round. The Skill identifies Reuters as the originator, grades it B2 with a rationale, and re-grades to B1 once an independent filing confirms the figure. ## Quick Start Ask the agent to grade a specific source on the Admiralty scale, for example: "Give this Reuters report an Admiralty rating — what is its A1–F6 grade and why?"
