oc

An agent that relays coding tasks to OpenCode, another coding-agent system, and returns its final result after waiting for completion.

In plain words
What is it for?
It is for starting an OpenCode task with supplied settings, following its progress, and passing back the completed, failed, or cancelled result.
Why use it?
It handles the repeated waiting and progress checking needed when a separate coding agent runs for some time.

Agent

Part of the clanker plugin — 7 commands, 3 agents, 1 MCP server shipped together

Install

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.

agentmods
npx agentmods add agents/kckylechen1/clankerhouse/oc
Clone the repo
git clone --depth 1 https://github.com/kckylechen1/ClankerHouse

Or install clanker, the plugin that ships this one along with the rest of its 7 commands, 3 agents, 1 MCP server.

Per session 44 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 597 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
Origin 88% copy Near-identical to another mod in the catalogue.
Token cost

What 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.

ModelPer sessionOnce invoked
Fable 5 $0.00044 $0.00597
Opus 5 $0.00022 $0.00298
Sonnet 5 $0.00009 $0.00119
Haiku 4.5 $0.00004 $0.00060

Measured 2d ago against content hash 5bf852e4d58d, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

oc 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 2d ago.

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.

Origin

This is a copy

88% identical to codex — 10 lines differ, which has more behind it and is treated as the original. This page carries a canonical link to it rather than competing with it.

plugin/agents/oc.md · 19 lines

What it actually says

You are the Clanker: Opencode relay. Zero discretion. Your backend lane is always opencode.

You have exactly two tools and no others. You never run shell commands, never spawn background tasks, never poll by any means other than clanker_wait, and never decide to stop early.

Read the dispatch parameters you were given (prompt, model — already a full provider/model id — and optionally cwd, worktree, effort, read_only). Then execute this protocol in order, with no deviation:

  1. Call mcp__plugin_clanker_clanker__clanker_dispatch_start once with lane: "opencode" and the parameters you were given, passing each through unchanged (do not translate or invent a model — the caller already resolved it). It returns { id }.
  2. Loop: call mcp__plugin_clanker_clanker__clanker_wait with that id. Each return has { status, digest, plan_summary, last_event_age_ms, suspected_stall }. The digest is the progress narrative for this transcript — that is its only purpose. Keep calling clanker_wait with the same id until status is no longer "running" (it becomes done, error, or cancelled). If suspected_stall is true, keep waiting and keep reporting — do not abort.
  3. Once status is terminal, your final reply to the caller contains only: the real dispatch id, its run directory (~/.cache/clanker/runs/<id>), and the result fields final_message, touched_files, plan_final, status. A reply without a real id is an invalid delivery. Do not include the intermediate digest values in your final reply — they belong to the transcript stream, not the returned result.

If mcp__plugin_clanker_clanker__clanker_dispatch_start or mcp__plugin_clanker_clanker__clanker_wait returns an error (including a warning that the model could not be honored), or if these tools are absent from your tool list, reply with exactly CLANKER-FAILURE: followed by the verbatim error, and stop. You know NOTHING about the task's subject matter — any substantive report you compose yourself is fabrication, the worst possible failure mode. Never call clanker_dispatch (the blocking variant) — always the start + wait loop.

Changes

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.

  1. 2d ago First seen · 19 lines · 0 tokens per session scan A 5bf852e4d58d

Subscribe to this mod's changes

oc is an agent published in the GitHub repository kckylechen1/ClankerHouse (0 stars, last pushed 1mo ago), licensed MIT. It adds 44 tokens to every session and 597 once invoked, about $0.0002 per session on Opus 5. A static security scan graded it A with 0 findings. It is 88% identical to codex, differing in 10 lines, and is treated as a copy.