structured-autonomy-plan

structured-autonomy-plan is an agent skill from the github/awesome-copilot collection that is a planning framework for breaking a development request into testable, commit-sized steps. It assumes the whole change will land in one pull request on its own branch, and it defines the plan as a series of steps that each correspond to a single commit. The first step is a research phase that is mandatory: the agent gathers the code context, the documentation, the dependencies and the patterns already used in the codebase before it plans anything, and it returns those findings before any planning begins. During this stage it writes no code. The second step determines the commits, breaking the request into an ordered sequence in which each commit can be tested on its own. It is useful for teams that review changes commit by commit and want each step to be small and verifiable. The plan is a proposal for the team to review, and its research depends on what the agent can read, so the parts it could not examine should be checked by a person. The skill is installed with the skills CLI from the github/awesome-copilot repository. Its skills.sh page showed about 8.8K installs and passing results from three security audits at the time of the scan.…