feat: externalize apps/analysis to Gitea repos, add analysis table

- Migration 007: repo_url on apps table + analysis table with FTS5
- Analysis struct, parser, CRUD, validation, hash computation
- Selective purge: remote-only apps/analysis preserved across fn index
- CLI: fn app list/clone/pull, fn analysis list/clone/pull
- search/show/list now include analysis results
- Apps removed from git tracking (content lives in Gitea repos)
- .gitkeep for apps/ and analysis/ dirs
- Bash functions: jupyter analysis pipeline, shell utilities
- Browser domain: CDP functions moved from infra to browser

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

bash/functions/shell/run_steps.md

@@ -0,0 +1,72 @@
+---
+name: run_steps
+kind: function
+lang: bash
+domain: shell
+version: "1.0.0"
+purity: impure
+signature: "run_steps(yaml_file: string, [--strict]) -> string"
+description: "Ejecuta pasos de un YAML generico donde cada step tiene action=command. Lee el YAML con yq, ejecuta cada paso secuencialmente con timeout configurable, captura exit code y output, respeta continue_on_error, y al final reporta JSON estandar a stdout via report_execution_json. Sale con exit code 0 (success), 1 (failure) o 2 (partial). Con --strict mapea partial a failure."
+tags: [execution, yaml, runner, standard, pipeline]
+uses_functions: [exit_with_status_bash_shell, report_execution_json_bash_shell]
+uses_types: []
+returns: []
+returns_optional: false
+error_type: "error_go_core"
+imports: []
+tested: false
+tests: []
+test_file_path: ""
+file_path: "bash/functions/shell/run_steps.sh"
+---
+
+## Ejemplo
+
+```yaml
+# /tmp/test_flow.yaml
+name: test_flow
+steps:
+  - name: check
+    action: command
+    command: "echo hello"
+  - name: list
+    action: command
+    command: "ls /tmp"
+    continue_on_error: true
+  - name: slow_step
+    action: command
+    command: "sleep 2"
+    timeout_ms: 1000
+    continue_on_error: true
+```
+
+```bash
+source bash/functions/shell/run_steps.sh
+
+run_steps /tmp/test_flow.yaml
+# Imprime JSON a stdout:
+# {
+#   "name": "test_flow",
+#   "status": "partial",
+#   "exit_code": 2,
+#   "started_at": "2026-04-01T10:00:00Z",
+#   "ended_at": "2026-04-01T10:00:01Z",
+#   "duration_ms": 1250,
+#   "steps_total": 3,
+#   "steps_ok": 2,
+#   "steps_failed": 1,
+#   "steps": [
+#     {"name": "check", "action": "command", "status": "ok", "elapsed_ms": 10, "output": "hello"},
+#     {"name": "list", "action": "command", "status": "ok", "elapsed_ms": 15},
+#     {"name": "slow_step", "action": "command", "status": "error", "elapsed_ms": 1001, "error": "timeout: comando excedio 1000ms"}
+#   ]
+# }
+# Sale con exit code 2 (partial)
+
+run_steps /tmp/test_flow.yaml --strict
+# Igual pero status="failure" y exit code=1
+```
+
+## Notas
+
+Requiere `yq` (v4+, modo expression `-e`) y `jq` (o la implementacion printf de report_execution_json). Solo soporta `action: command` — otros actions se marcan como error del paso. El campo `timeout_ms` por defecto es 30000 (30s); si el comando excede el timeout, `timeout(1)` sale con exit code 124 y se registra el error correspondiente. El output del comando (stdout+stderr combinados) se sanitiza reemplazando tabuladores y saltos de linea por espacios antes de escribir al TSV temporal. Las funciones `exit_with_status` y `report_execution_json` se cargan via `source` desde el mismo directorio que `run_steps.sh`.