Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx agentmods add commands/openclaw/crabbox/attachgit clone --depth 1 https://github.com/openclaw/crabboxWhat it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5 | $0.00000 | $0.00747 |
| Opus 5 | $0.00000 | $0.00374 |
| Sonnet 5 | $0.00000 | $0.00149 |
| Haiku 4.5 | $0.00000 | $0.00075 |
Grade A, and why
attach scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured yesterday.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 82 lines — stays where its author put it; the contents beside it link to each section on GitHub.
attach
crabbox attach follows the recorded events of an active coordinator run and
prints them as they arrive, exiting when the run finishes.
crabbox attach run_abcdef123456
crabbox attach --id run_abcdef123456 --after 42
crabbox attach run_abcdef123456 --poll 500ms
The run id may be passed as a positional argument or via --id; both forms are
equivalent. Run ids look like run_<hex>.
How it works
attach first tries the authenticated coordinator control WebSocket, which
streams run events live. If the socket cannot be dialed, fails to subscribe, or
errors mid-stream before the run finishes, attach falls back to the HTTPS
events API and keeps reading from the last sequence number it printed. Either
way it advances a running cursor so no event is shown twice.
Events are routed by stream:
stdoutandstderrpreview events are written back to stdout and stderr respectively, preserving the original stream split;- all other events (lease, bootstrap, sync, command-start, command-finish, release, and similar lifecycle events) are printed to stderr as a single line carrying the sequence number, event type, phase, timestamp, and message.
When the WebSocket falls behind the run's current sequence, attach drains the
backlog in bounded pages, re-subscribing after each full page, before waiting on
live pushes. When the run is already finished, attach prints any remaining
events and exits. When the run is still active, it waits for streamed events or
polls until it observes the run leave the running state.
attach is a viewer, not detached execution. It follows the events the original
CLI emitted while running the command; it does not own or control that command.
If the original CLI process dies, the run remains inspectable through
history, events, and logs, but attach
cannot resume or restart it.
Bounded output
Output preview events are a bounded view of the command output. The coordinator stores run logs in 64 KiB chunks up to an 8 MiB cap and marks the run truncated when that cap is exceeded. For the full retained output after a run completes, use logs.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- yesterday First seen · 82 lines · 0 tokens per session scan A 775e4346f8e9
attach is a command published in the GitHub repository openclaw/crabbox (1,343 stars, last pushed yesterday), licensed MIT. It costs nothing until one of its globs matches a file; then it loads 747 tokens. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other commands, from other repositories
checklist
Generate a custom checklist for the current feature based on user requirements.
clarify
Identify underspecified areas in the current feature spec by asking up to 5 highly targeted clarification questions and encoding answers back into the spec.
specify
Create or update the feature specification from a natural language feature description.
analyze
Perform a non-destructive cross-artifact consistency and quality analysis across spec.md, plan.md, and tasks.md after task generation.
converge
Assess the current codebase against the feature's spec, plan, and tasks, then append any remaining unbuilt work as new tasks to tasks.md so implement can complete it.
implement
Execute the implementation plan by processing and executing all tasks defined in tasks.md.