| name | flex:archeology |
| description | Reconstruct project history from Flex evidence. Use when the user wants to recover workstreams, shifts, session assignments, artifact lineages, file history, or other retroactive knowledge surfaces from AI coding-agent sessions, git/file activity, context docs, and Flex cells. |
| user-invocable | true |
| argument-hint | workflow + target, e.g. 'workstreams for <YOUR_PROJECT>/context into <YOUR_PROJECT>/workstreams-v2' |
flex:archeology
Flex archeology turns historical evidence into durable project memory. The sources are AI coding-agent sessions, git/file activity, context docs, changelogs, plans, commits, and Flex search surfaces. The output is not a speculative plan; it is a reconstructed map of what the project has actually been doing.
Router
Use the smallest workflow that fits:
- Workstreams: read
workstreams/workflow.md when reconstructing workstreams, shifts, timelines, and session assignments for a project.
- File history: read
file-history/workflow.md when tracing recoverable versions of one file.
- Artifact lineage: read
artifact-lineage/workflow.md when writing changelog-style lineage for one artifact.
If the user says only "archeology" and names a project, default to the workstreams workflow. If they name a specific file, default to file history or artifact lineage depending on whether they want versions or narrative.
Core Rules
- Prefer evidence over taxonomy.
- Reference durable project docs before applying recipe text.
- Keep the artifact compact and agent-useful.
- Make uncertain assignments visible instead of forcing them.
- A session may belong to multiple workstreams.
- Do not create query files unless the user explicitly asks for reusable query artifacts.
