feat: add harness skill with hooks install/uninstall support (#156)

Add multi-session autonomous agent harness with progress checkpointing,
failure recovery, task dependencies, and post-completion self-reflection.

- Add harness module to config.json (copy_dir with hooks.json)
- Add 7 hook scripts: stop, sessionstart, teammateidle, subagentstop,
  claim, renew, self-reflect-stop + shared _harness_common.py
- Fix self-reflect-stop: only triggers when harness was initialized
  (checks harness-tasks.json existence), not on every session
- Add unmerge_hooks_from_settings() to uninstall.py for clean hook removal
- Add unit tests (57 tests) and E2E test (100 tasks + 5 self-reflect)

Generated with SWE-Agent.ai

Co-Authored-By: SWE-Agent.ai <noreply@swe-agent.ai>

config.json

@@ -196,6 +196,18 @@
         }
       ]
     },
+    "harness": {
+      "enabled": false,
+      "description": "Multi-session autonomous agent harness with progress checkpointing, failure recovery, task dependencies, and post-completion self-reflection",
+      "operations": [
+        {
+          "type": "copy_dir",
+          "source": "skills/harness",
+          "target": "skills/harness",
+          "description": "Install harness skill with hooks (Stop, SessionStart, TeammateIdle, SubagentStop, self-reflect)"
+        }
+      ]
+    },
     "claudekit": {
       "enabled": false,
       "description": "ClaudeKit workflow: skills/do + global hooks (pre-bash, inject-spec, log-prompt)",