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/greenpolo/cc-multi-cli-plugin/rescuegit clone --depth 1 https://github.com/greenpolo/cc-multi-cli-pluginWhat 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 | $0.00017 | $0.00460 |
| Opus 5 | $0.00009 | $0.00230 |
| Sonnet 5 | $0.00003 | $0.00092 |
| Haiku 4.5 | $0.00002 | $0.00046 |
Grade A, and why
rescue 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.
What it actually says
Invoke the multi:codex-rescue subagent via the Agent tool, forwarding the user's request as the prompt.
The subagent runs Codex in open-ended rescue mode — hand it a stuck problem, a confusing bug, or a substantial unbounded task and it investigates independently.
This is distinct from /codex:execute (structured execution of a clear plan step) and /codex:review (read-only review). Use /codex:rescue when Claude is stuck, wants a second opinion, or needs a deeper root-cause investigation.
Raw user request: $ARGUMENTS
Execution:
- For a small, clearly bounded rescue, invoke the subagent in the foreground (blocking) — the result comes back inline.
- For an open-ended, multi-step, or long-running investigation, invoke the subagent with the
Agenttool'srun_in_background: true. The subagent runs Codex to completion in the foreground, so the harness wakes you with a<task-notification>the moment Codex finishes, errors, OR the inactivity watchdog kills a hang — even after this turn ends. This is the path that surfaces failures instead of stalling silently. - Do NOT add the companion's
--backgroundflag to get this — that detaches a worker the harness cannot see (no notification). Reserve--backgroundfor explicit fire-and-forget the user wants to poll via/multi:status. - Preserve
--model,--effort,--resume,--freshfor the forwarded command — the subagent reads them. - If the user passes
--resume, the subagent continues the latest Codex thread for this repo; otherwise it starts fresh. - If the request includes no prompt text, ask what Codex should investigate before proceeding.
Return Codex's output verbatim once the subagent completes.
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 · 26 lines · 17 tokens per session scan A ad531aec108d
rescue is a command published in the GitHub repository greenpolo/cc-multi-cli-plugin (89 stars, last pushed 15d ago), licensed Apache-2.0. It adds 17 tokens to every session and 460 once invoked, about $0.0001 per session on Opus 5. 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
delegate
Delegate to Cursor, Composer, or cursor-agent via cursor-delegate-mcp (see delegate skill).
icpg-bootstrap
Infer ReasonNodes from existing git commit history. One-time setup for existing codebases.
version
Display current guide and Claude Code versions.
go-review
Go code review for idiomatic patterns.
release-harn
Run the tag-first Harn release workflow.
esp-harden
Harden and inspect ESP32 firmware for field failures, crashes, memory, and security.