Synthesizes conversations into product INTENT.md maps or committed effort intake files.
npx skills add https://github.com/peachest/skills --skill to-intent-peachest
Please help me install this Agent Skill. Skill: to-intent Source: https://github.com/peachest/skills/tree/main/in-progress/to-intent Command: npx skills add https://github.com/peachest/skills --skill to-intent-peachest
High-intent search queries and answers about installing and using this skill.
Run the Skill in product mode to synthesize the conversation into an INTENT.md at the repository root. It drafts the product brief, releases, and stories with priorities and Done-when criteria, then quizzes you for approval before writing the file.
Use effort mode, which writes a committed intent/<slug>.md file with Problem, Proposed outcome, Constraints, and Open questions sections. The frontmatter records the origin (story id, idea, incident, diagnosis, or ticket) and a status lifecycle from draft to done.
Yes, a later run acts as an update pass: it re-reads INTENT.md, re-synthesizes against the new conversation, keeps ids of surviving stories, and drops completed ones while telling you which. Story ids are stable forever unless a story is replaced.
A story whose Done-when cannot be sharpened into a single verifiable statement is treated as fog, not a story. It is parked in a later release at could priority or handed to a decision-mapping workflow whose tickets clear the fog first.
Use effort mode when the conversation concerns one change, fix, or incident response rather than the product as a whole. Product mode is the default for discussions about releases, story cuts, and priorities across the whole product.
Neuron-Mr-WhiteGuide collaborative brainstorming sessions into structured design specifications.
kinncjConvert feature ideas into Gherkin story files with frontmatter status tracking.
SethGammonConverts loose ideas into git-tracked decision maps of typed investigation tickets resolved one per session.
ssccioCreate structured PRD and progress log files for ralph-loop execution.
