What problem does it solve? AI agents hold large amounts of conversation, investigation, and execution context, and dumping that raw context into persistent text produces documents full of scaffolding, chronology, and references only the author understands. This Skill enforces a disciplined writing pipeline so documentation, ADRs, Issues, PRs, commit messages, and comments stand alone as reader-oriented artifacts. ## Core Features & Use Cases - Three-phase writing pipeline: Select communicative content for the audience, Compose standalone prose, then Reread the draft as a reader without the original context. - Artifact-specific guidance: Tailored rules for documentation, ADRs, GitHub Issues, Pull Requests, commit messages, code comments, and review comments. - Temporal information policy: Clear criteria for when version, migration, or history details belong in an artifact versus when they are context leakage. - Use Case: After finishing a debugging session, use this Skill to write the Pull Request description so reviewers see the intent, design decisions, and validation rather than a transcript of trial-and-error. ## Quick Start Use the writing-discipline skill to draft the commit message and PR description for the changes I just made.
