/observe
Agents Observe dashboard and server management.
Usage
/observe view — Open the current session in the dashboard/observe stats — Open the current session's stats modal in the dashboard/observe — Open the dashboard URL/observe status — Show server health and config details/observe start — Start the server/observe stop — Stop the server/observe restart — Restart the server/observe logs-server — Show recent Docker container logs/observe logs-cli — Tail the local cli.log file/observe logs-mcp — Tail the local mcp.log file/observe debug — Diagnose server issues (health, docker logs, mcp.log, cli.log)
Instructions
All commands are invoked via the wrapper at scripts/cli.sh, which resolves the path to observe_cli.mjs relative to this skill's install location. Do not reach outside the skill dir with ../ paths — the agentskills spec disallows it and it breaks portability across agents.
The subcommand is in $ARGUMENTS. If empty, default to showing the dashboard URL.
/observe view
Opens the current session in the dashboard.
- Run health to get the dashboard origin:
scripts/cli.sh health
- From the output, take the
Dashboard: URL (e.g. http://localhost:4981). If exit code 1, the server isn't running — tell the user to run /observe start and stop here.
- Construct the session URL:
<dashboard>/#/_/${CLAUDE_SESSION_ID} (the _ is the project placeholder; the dashboard resolves the real project from the session id).
- Open it in the user's default browser using the platform-appropriate command —
open <url> on macOS, xdg-open <url> on Linux, start <url> on Windows. Pick based on the Platform: line in your environment context.
- Also print the URL in your response so the user can re-open it if needed.
/observe stats
Opens the current session's stats modal in the dashboard, using a deep-link URL.
- Run health to get the dashboard origin:
scripts/cli.sh health
- From the output, take the
Dashboard: URL (e.g. http://localhost:4981). If exit code 1, the server isn't running — tell the user to run /observe start and stop here.
- Construct the deep link:
<dashboard>/#/_/${CLAUDE_SESSION_ID}:session.stats
- Open it in the user's default browser using the platform-appropriate command —
open <url> on macOS, xdg-open <url> on Linux, start <url> on Windows. Pick based on the Platform: line in your environment context.
- Also print the URL in your response so the user can re-open it if needed.
/observe (no args)
- Run:
scripts/cli.sh health
- If exit code 0: show the dashboard URL from the output.
- If exit code 1: tell the user the server is not running and suggest
/observe start or /observe status.
/observe status
- Run:
scripts/cli.sh health
- Show the full output to the user (includes version, runtime, ports, log paths).
- If the output contains "Version mismatch", tell the user and offer
/observe restart.
- If exit code 1, show the output and suggest
/observe start.
/observe start
- Run:
scripts/cli.sh start
- Show the output to the user. If successful, include the dashboard URL.
/observe stop
- Run:
scripts/cli.sh stop
- Confirm to the user that the server has been stopped.
/observe restart
- Run:
scripts/cli.sh restart
- Show the output to the user. If successful, include the dashboard URL.
/observe logs-server
- Run:
scripts/cli.sh logs-server -n 50
- Show the output to the user. Do NOT use
-f (follow) — it would hang.
/observe logs-cli
- Run:
scripts/cli.sh logs-cli -n 50
- Show the output to the user.
/observe logs-mcp
- Run:
scripts/cli.sh logs-mcp -n 50
- Show the output to the user.
/observe debug
Run these checks in sequence. Read each output before running the next — use what you learn to diagnose the issue.
Server health:
scripts/cli.sh health
Docker container logs (last 20 lines):
scripts/cli.sh logs-server -n 20
MCP log (last 20 lines):
scripts/cli.sh logs-mcp -n 20
CLI log (last 20 lines):
scripts/cli.sh logs-cli -n 20
Analyze the results and tell the user:
- Is the server running? What version?
- Are there errors in the docker logs? (look for crash loops, port conflicts, DB errors)
- Are there errors in mcp.log? (look for startServer failures, image pull errors)
- Are there errors in cli.log? (look for ECONNREFUSED, hook delivery failures)
- Suggest specific fixes based on what you find.
1---2name: observe3description: Agents Observe dashboard and server management4---56# /observe78Agents Observe dashboard and server management.910## Usage1112- `/observe view` — Open the current session in the dashboard13- `/observe stats` — Open the current session's stats modal in the dashboard14- `/observe` — Open the dashboard URL15- `/observe status` — Show server health and config details16- `/observe start` — Start the server17- `/observe stop` — Stop the server18- `/observe restart` — Restart the server19- `/observe logs-server` — Show recent Docker container logs20- `/observe logs-cli` — Tail the local cli.log file21- `/observe logs-mcp` — Tail the local mcp.log file22- `/observe debug` — Diagnose server issues (health, docker logs, mcp.log, cli.log)2324## Instructions2526All commands are invoked via the wrapper at `scripts/cli.sh`, which resolves the path to `observe_cli.mjs` relative to this skill's install location. Do not reach outside the skill dir with `../` paths — the agentskills spec disallows it and it breaks portability across agents.2728The subcommand is in `$ARGUMENTS`. If empty, default to showing the dashboard URL.2930### /observe view3132Opens the current session in the dashboard.33341. Run health to get the dashboard origin:35 ```bash36 scripts/cli.sh health37 ```382. From the output, take the `Dashboard:` URL (e.g. `http://localhost:4981`). If exit code 1, the server isn't running — tell the user to run `/observe start` and stop here.393. Construct the session URL: `<dashboard>/#/_/${CLAUDE_SESSION_ID}` (the `_` is the project placeholder; the dashboard resolves the real project from the session id).404. Open it in the user's default browser using the platform-appropriate command — `open <url>` on macOS, `xdg-open <url>` on Linux, `start <url>` on Windows. Pick based on the `Platform:` line in your environment context.415. Also print the URL in your response so the user can re-open it if needed.4243### /observe stats4445Opens the current session's stats modal in the dashboard, using a deep-link URL.46471. Run health to get the dashboard origin:48 ```bash49 scripts/cli.sh health50 ```512. From the output, take the `Dashboard:` URL (e.g. `http://localhost:4981`). If exit code 1, the server isn't running — tell the user to run `/observe start` and stop here.523. Construct the deep link: `<dashboard>/#/_/${CLAUDE_SESSION_ID}:session.stats`534. Open it in the user's default browser using the platform-appropriate command — `open <url>` on macOS, `xdg-open <url>` on Linux, `start <url>` on Windows. Pick based on the `Platform:` line in your environment context.545. Also print the URL in your response so the user can re-open it if needed.5556### /observe (no args)57581. Run:59 ```bash60 scripts/cli.sh health61 ```622. If exit code 0: show the dashboard URL from the output.633. If exit code 1: tell the user the server is not running and suggest `/observe start` or `/observe status`.6465### /observe status66671. Run:68 ```bash69 scripts/cli.sh health70 ```712. Show the full output to the user (includes version, runtime, ports, log paths).723. If the output contains "Version mismatch", tell the user and offer `/observe restart`.734. If exit code 1, show the output and suggest `/observe start`.7475### /observe start76771. Run:78 ```bash79 scripts/cli.sh start80 ```812. Show the output to the user. If successful, include the dashboard URL.8283### /observe stop84851. Run:86 ```bash87 scripts/cli.sh stop88 ```892. Confirm to the user that the server has been stopped.9091### /observe restart92931. Run:94 ```bash95 scripts/cli.sh restart96 ```972. Show the output to the user. If successful, include the dashboard URL.9899### /observe logs-server1001011. Run:102 ```bash103 scripts/cli.sh logs-server -n 50104 ```1052. Show the output to the user. Do NOT use `-f` (follow) — it would hang.106107### /observe logs-cli1081091. Run:110 ```bash111 scripts/cli.sh logs-cli -n 50112 ```1132. Show the output to the user.114115### /observe logs-mcp1161171. Run:118 ```bash119 scripts/cli.sh logs-mcp -n 50120 ```1212. Show the output to the user.122123### /observe debug124125Run these checks in sequence. Read each output before running the next — use what you learn to diagnose the issue.1261271. **Server health:**128 ```bash129 scripts/cli.sh health130 ```1311322. **Docker container logs (last 20 lines):**133 ```bash134 scripts/cli.sh logs-server -n 20135 ```1361373. **MCP log (last 20 lines):**138 ```bash139 scripts/cli.sh logs-mcp -n 20140 ```1411424. **CLI log (last 20 lines):**143 ```bash144 scripts/cli.sh logs-cli -n 20145 ```1461475. **Analyze the results** and tell the user:148 - Is the server running? What version?149 - Are there errors in the docker logs? (look for crash loops, port conflicts, DB errors)150 - Are there errors in mcp.log? (look for startServer failures, image pull errors)151 - Are there errors in cli.log? (look for ECONNREFUSED, hook delivery failures)152 - Suggest specific fixes based on what you find.
simple10