Repository navigation
feat: a guide for AI agents, and JSON from init and apply - #16
Merged
Merged
Conversation
`harness2 guide` prints how an AI agent should drive harness2 on a person's behalf: ask before spending, bound runs with --max-calls, read --json, answer prompts with flags, never edit runs or harness files directly, the init/task/doctor/run --detach/status/stop/diff/eval/apply workflow, where things are stored and what the exit codes mean, followed by the current project's settings, tasks and runs. `guide --install claude` installs the same text as a user-level Claude Code skill, which harness2's own runs never load. `harness2 chat` now builds on it. `init --json` and `apply --json` print a result object on stdout and keep their human-readable text on stderr; `apply --json` refuses to ask for confirmation and needs --yes or --dry-run. Co-Authored-By: Claude Opus 5.5 <[email protected]>
This was referenced Oct 19, 2026
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
More and more, a CLI like this is driven by an AI agent on someone's behalf. This PR is the agent-facing half. The other half is the run-control PR:
run --detach,stop, and JSON errors.harness2 guideprints the contract an agent should follow, then the project's current settings, tasks and runs. The contract:--max-calls--jsonoutput, and answer prompts with flagsruns/or the harness directlyrun --detach→status --json/stop→diff/eval→apply --yesguide --install claudeinstalls the guide as a user-level Claude Code skill. harness2's own runs never load user skills, so it can't contaminate a run.harness2 chatreuses the same guide, so it lives in one place.init --jsonandapply --jsonprint a result object on stdout and keep their human-readable text on stderr.apply --jsonnever prompts: it requires--yesor--dry-run.Test plan
harness2 guidechecked by hand, inside and outside a project.ruff🤖 Generated with Claude Code