Scenario 4: Scan an existing repository and import its current harness setup
Use this when a repository already contains agent instructions, skills, MCP config, rules, hooks, or agent files that you want HarnessTap to manage.
Frequency: Common · Status: Shipped
Use this when a repository already contains agent instructions, skills, MCP config, rules, hooks, or agent files that you want HarnessTap to manage.
Typical commands:
harnesstap scan . --dry-run # preview what would be imported
harnesstap scan .
harnesstap scan . --harness claude-code # scope to a single harness
harnesstap resource list
harnesstap resource show <resource-name-or-id>This is the normal "adopt an existing repo" workflow. It works well for repos
that already contain files such as CLAUDE.md, AGENTS.md, .claude/skills/,
.cursor/rules/, .codex/config.toml, and similar harness-specific layouts.
Re-scanning is idempotent within a run: resources deduplicate by type:name
before insertion, so running the same scan twice does not create duplicate
records.
Scenario 3: Override harness preferences for one repository
Use this when your global preference is not the right fit for a specific repo. For example, you may usually author in Claude Code but want one project to...
Scenario 5: Build a reusable layer from imported resources
Use this when you want to turn a working repository setup into a named, reusable harness baseline.