A snapshot: read at commit 707a04c7ce8 on 2026-10-18T00:06:01Z by majordomus 0.29.0. Nothing on this page is read again. See it live
use-case
49 of the 49 object(s) of kind use-case. Each has an address of its own, derived from its identity.
| Identity | Title | Source |
|---|---|---|
| accept-or-refuse-finished-work | Decide whether finished work is actually finished | .ai/repo/use-cases/accept-or-refuse-finished-work.md |
| add-a-use-case-and-prove-it | Add a use case and let the tool prove it | .ai/repo/use-cases/add-a-use-case-and-prove-it.md |
| adopt-an-existing-repository | Adopt a repository that already has its own rules | .ai/repo/use-cases/adopt-an-existing-repository.md |
| block-acceptance-on-an-open-question | Block acceptance on a question nobody has answered | .ai/repo/use-cases/block-acceptance-on-an-open-question.md |
| capture-the-prompts-that-started-the-work | Keep the prompts that started the work, below the model rather than around it | .ai/repo/use-cases/capture-the-prompts-that-started-the-work.md |
| carry-a-blocker-across-a-handover | Hand work over with a question still open, and keep it blocking | .ai/repo/use-cases/carry-a-blocker-across-a-handover.md |
| catch-an-agent-that-says-it-is-done | Catch an agent that says it is done when it is not | .ai/repo/use-cases/catch-an-agent-that-says-it-is-done.md |
| checkpoint-long-work | Checkpoint long work so a stop costs minutes, not the day | .ai/repo/use-cases/checkpoint-long-work.md |
| classify-what-belongs-in-the-context | Decide what belongs in the AI context, as data | .ai/repo/use-cases/classify-what-belongs-in-the-context.md |
| complete-an-issue-only-with-its-evidence | Complete an issue only when its evidence exists | .ai/repo/use-cases/complete-an-issue-only-with-its-evidence.md |
| deliver-issues-in-waves | Deliver issues in waves the graph computes, and take the next one | .ai/repo/use-cases/deliver-issues-in-waves.md |
| document-every-directory-of-the-layer | Refuse a directory of the layer that documents nothing | .ai/repo/use-cases/document-every-directory-of-the-layer.md |
| extend-what-the-executable-serves | Extend what the executable serves by adding a file, never by rebuilding it | .ai/repo/use-cases/extend-what-the-executable-serves.md |
| find-an-object-without-reading-everything | Find an object of the layer without reading everything | .ai/repo/use-cases/find-an-object-without-reading-everything.md |
| find-out-what-drifted | Find out what has drifted since anyone last looked | .ai/repo/use-cases/find-out-what-drifted.md |
| follow-a-skill-the-repository-defines | Follow a procedure the repository defines, and add another one | .ai/repo/use-cases/follow-a-skill-the-repository-defines.md |
| gate-ci-on-the-tool-itself | Gate CI on the tool itself, with an exit code that is a contract | .ai/repo/use-cases/gate-ci-on-the-tool-itself.md |
| hand-the-repository-to-a-reader | Hand the repository to someone who cannot clone it | .ai/repo/use-cases/hand-the-repository-to-a-reader.md |
| hand-work-between-sessions | Hand unfinished work to the next session | .ai/repo/use-cases/hand-work-between-sessions.md |
| install-the-tool-in-one-command | Install the tool in one command, on a machine with no toolchain | .ai/repo/use-cases/install-the-tool-in-one-command.md |
| keep-decisions-out-of-the-transcript | Record a decision so the next worker can find it | .ai/repo/use-cases/keep-decisions-out-of-the-transcript.md |
| keep-the-bootstrap-thin-and-within-budget | Keep every provider bootstrap thin, generated and within budget | .ai/repo/use-cases/keep-the-bootstrap-thin-and-within-budget.md |
| keep-what-an-episode-learned | Keep what an episode learned, without a model and without a conversation | .ai/repo/use-cases/keep-what-an-episode-learned.md |
| know-whether-this-work-is-finished | Know whether the work in this checkout is finished | .ai/repo/use-cases/know-whether-this-work-is-finished.md |
| know-which-tool-is-running | Check the version before trusting a diagnosis | .ai/repo/use-cases/know-which-tool-is-running.md |
| let-the-provider-draw-the-episode-boundary | Have the session opened and closed without anybody remembering to | .ai/repo/use-cases/let-the-provider-draw-the-episode-boundary.md |
| migrate-from-the-old-layout | Move a repository from the pre-.ai layout to the layer | .ai/repo/use-cases/migrate-from-the-old-layout.md |
| open-and-close-a-session | Record what one worker did in one sitting | .ai/repo/use-cases/open-and-close-a-session.md |
| plan-a-commit-and-judge-its-message | Divide a working tree into commits, and hold every message to a policy | .ai/repo/use-cases/plan-a-commit-and-judge-its-message.md |
| plan-the-work-as-data | Keep the plan as milestones and issues the tool can validate | .ai/repo/use-cases/plan-the-work-as-data.md |
| prove-a-rule-is-enforced | Prove a rule is actually enforced, not merely written down | .ai/repo/use-cases/prove-a-rule-is-enforced.md |
| prove-performance-with-benchmarks | Prove the tool is fast, not assume it, and refuse a regression | .ai/repo/use-cases/prove-performance-with-benchmarks.md |
| prove-what-a-task-owed-before-calling-it-done | Prove what a task owed before calling it done | .ai/repo/use-cases/prove-what-a-task-owed-before-calling-it-done.md |
| read-back-what-happened | Read back what happened, and keep the ledger within its cap | .ai/repo/use-cases/read-back-what-happened.md |
| read-only-the-context-that-fits | Read only the context that fits the budget, and know what was dropped | .ai/repo/use-cases/read-only-the-context-that-fits.md |
| read-the-rules-the-tool-applies | See which rules apply here and what enforces them | .ai/repo/use-cases/read-the-rules-the-tool-applies.md |
| record-a-decision-before-it-is-forgotten | Record a decision as data, and prove the tool cannot accept it for you | .ai/repo/use-cases/record-a-decision-before-it-is-forgotten.md |
| recover-an-episode-nobody-closed | Close an episode whose client never came back | .ai/repo/use-cases/recover-an-episode-nobody-closed.md |
| resume-from-a-prompt-asset | Start a session from a rendered framing, not a pasted transcript | .ai/repo/use-cases/resume-from-a-prompt-asset.md |
| resume-in-the-right-worktree | Resume in the right worktree, never against somebody else’s task | .ai/repo/use-cases/resume-in-the-right-worktree.md |
| run-several-workers-at-once | Run several workers in one repository without them colliding | .ai/repo/use-cases/run-several-workers-at-once.md |
| say-what-reached-the-remote-and-what-did-not | Say what reached the remote and what did not | .ai/repo/use-cases/say-what-reached-the-remote-and-what-did-not.md |
| say-what-the-tool-is-for | Say what your work is for, once, and have every surface say it | .ai/repo/use-cases/say-what-the-tool-is-for.md |
| see-what-the-repository-holds-without-reading-it | See what the repository holds, and run one of its capabilities, without reading a file | .ai/repo/use-cases/see-what-the-repository-holds-without-reading-it.md |
| serve-the-layer-to-ai-clients | Serve the repository's AI layer to every AI client through one shared server | .ai/repo/use-cases/serve-the-layer-to-ai-clients.md |
| trace-a-change-to-the-context-it-affects | Trace a change to the context documents it affects | .ai/repo/use-cases/trace-a-change-to-the-context-it-affects.md |
| trust-the-policy-before-reading-it | Trust the policy and the profiles only after they are validated | .ai/repo/use-cases/trust-the-policy-before-reading-it.md |
| use-the-tool-as-your-own-in-a-shared-repository | Use the tool as your own in a shared repository, and propose it to the team when it has made its case | .ai/repo/use-cases/use-the-tool-as-your-own-in-a-shared-repository.md |
| work-on-a-branch-in-its-canonical-worktree | Work on a branch in the one worktree it belongs to | .ai/repo/use-cases/work-on-a-branch-in-its-canonical-worktree.md |