Convert prose, rules, and workflows into valid YON v2.0 records.
npx skills add https://github.com/allemaar/open-skills --skill yon-write-allemaar
Please help me install this Agent Skill. Skill: yon-write Source: https://github.com/allemaar/open-skills/tree/main/skills/yon-write Command: npx skills add https://github.com/allemaar/open-skills --skill yon-write-allemaar
@RULE levels while preserving hedging and uncertainty verbatim. - Decision Gates: Structured selection of profile (core, decl, exec, audit, cognitive, agent), format (canon, min, ultra), and document kind before writing. - Validation Checklist: Built-in verification of @DOC placement, @BEGIN/@END pairing, @MAP quoting, and required fields per tag. - Use Case: Convert a messy prose API policy ("all calls must have auth tokens, premium users get 500/min") into a parseable kind=rule YON document with @RULE, @MAP, and @STAMP records. ## Quick Start Convert this rate limiting policy into a YON rule document using the decl profile and min format.High-intent search queries and answers about installing and using this skill.
Detect section boundaries from topic shifts, map constraint words to rule levels ("must" to MUST, "never" to MUST_NOT, "should" to SHOULD), move purpose statements to @INTENT, and drop conversational noise. Preserve hedging like "probably" verbatim rather than resolving it to confidence values.
@DOC requires ver, id, and title and must be the first non-comment record. @RULE requires lvl (MUST/MUST_NOT/SHOULD/SHOULD_NOT/MAY), when, and then. Other tags like @STEP need rid, n:int, and op; @MAP needs name and pairs with both sides quoted.
Avoid YON for API responses, config files, database records, simple key-value data, and messages under roughly 340 tokens, where JSON, TOML, or plain text are cheaper. YON fits system prompts, agent instructions, policies, workflows, and provenance tracking where intent and audit travel together.
Choose profile by content: decl for rules and schemas, exec for workflows, audit for provenance, cognitive for thought chains, agent for multi-agent systems, core for basic documents. Use fmt=min for LLM pipelines (the recommended default), canon for human-readable docs, and ultra for cost-critical transport.
Every @BEGIN must include mime= and a boundary of 8 or more characters, and each @END must repeat the same TAG and boundary exactly. When multiple blocks exist, each needs a unique id= or the parser raises a duplicate ID error. Blocks must not nest.
Innorve-IncEncode governance prose into a machine-enforceable YAML/JSON Policy Specification.
bswrundquistCommit staged changes using conventional commits and push them to the remote repository.
Syedyasir001Executes written implementation plans task-by-task with verification checkpoints and review gates.
winsorllcValidate, format, convert, query, flatten, unflatten, and merge JSON, YAML, and TOML data files.