build-acp-agent

A skill for building an ACP agent or server. ACP is a protocol that lets an editor or client communicate with a coding agent.

In plain words
What is it for?
It helps design, code, test, package, and submit ACP agents, including streaming updates and tool display.
Why use it?
It provides a defined implementation path for protocol negotiation, sessions, permissions, and authentication.

Skill for Claude CodeCodex

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 skills/gaelic-ghost/socket/build-acp-agent
Any agent
npx skills add gaelic-ghost/socket --skill build-acp-agent
Clone the repo
git clone --depth 1 https://github.com/gaelic-ghost/socket

Made for: Claude Code, Codex.

Per session 42 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 864 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found 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 $0.00042 $0.00864
Opus 5 $0.00021 $0.00432
Sonnet 5 $0.00008 $0.00173
Haiku 4.5 $0.00004 $0.00086

Measured 2d ago against content hash 847f28bff3c3, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

build-acp-agent 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.

plugins/agent-portability-skills/skills/build-acp-agent/SKILL.md · 87 lines

How it starts

The opening of the file, as written. The whole thing — 87 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Build ACP Agent

Build the agent process that an ACP client launches. Do not build an editor plug-in or MCP server when the requirement is an editor-native coding agent. Read references/acp-implementation-map.md before selecting an SDK or registry distribution.

Implementation Workflow

  1. Define concrete client targets and the user-visible behavior they require.
  2. Use ACP v1, the current latest protocol, and an official SDK when its language and runtime fit. Treat v2 and non-completed RFDs as experimental; do not ship them as the portable baseline. Use raw JSON-RPC only when an SDK cannot support a required platform or feature.
  3. Implement initialization first:
    • negotiate protocolVersion
    • advertise only implemented agent capabilities
    • inspect client filesystem and terminal capabilities before using them
    • advertise real authentication methods and implementation information
  4. Implement the required session baseline: new, prompt, cancel, and update.
  5. Add optional load, resume, fork, list, close, modes, configuration, plans, slash commands, rich prompt content, or MCP transports only with matching capability declarations and tests. Keep draft session/delete, logout, elicitation, additional directories, and remote transports behind explicit experimental flags and interoperability tests.
  6. Render tool calls, diffs, terminals, plans, and streamed content in the ACP structures the target clients understand.
  7. Route dangerous operations through permission requests and fail closed when approval transport fails or times out.
  8. Keep stdout exclusively for ACP JSON-RPC. Send logs and diagnostics to stderr without leaking credentials or prompt content unnecessarily.
  9. Bind every session to its client-provided cwd and any negotiated additional roots. Isolate cancellation, model choice, history, and mutable state by session.
  10. Test initialization, authentication, session lifecycle, streaming, cancellation, permissions, invalid messages, process shutdown, and each advertised optional capability against representative clients.

Read the full file on GitHub · 87 lines

Files

What ships with it

2 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.

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 · 87 lines · 42 tokens per session scan A 847f28bff3c3

Subscribe to this mod's changes

build-acp-agent is a skill published in the GitHub repository gaelic-ghost/socket (7 stars, last pushed 7d ago), licensed Apache-2.0. It adds 42 tokens to every session and 864 once invoked, about $0.0002 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 2026-08-31.

Related

Other skills, from other repositories

baocut

BaoCut-only operator for the installed bcut CLI and .bcut projects. Implicitly trigger only when the request explicitly names BaoCut or bcut, targets a .bcut project or BaoCut Subtitle Studio, or continues a BaoCut workflow already established in the conversation. Do not trigger solely for generic audio or video…

JimLiu/baocut · 160 tokens

xcode-disk-cleanup

Audit and safely clean Xcode-related developer storage on macOS. Use whenever a developer mentions low disk space, Xcode storage, DerivedData, simulator devices or runtimes, stale beta platforms in Xcode Settings, DeviceSupport, archives, dSYMs, CocoaPods caches, simulator dyld caches, project .build folders…

AvdLee/Xcode-Disk-Cleanup-Agent-Skill · 147 tokens

accelerate-local-work

Use AtomLane to compile and execute safe atomic parallel plans for macOS tasks containing composite shell or package scripts, Make or Compose graphs, native test/build runners, repeated independent work, dependency DAGs, or Apple-silicon operators. Use at task start or an execution boundary when structured local work…

cloudguo123/atomlane · 89 tokens

use-gemini-web

Use the user's logged-in Gemini Web session as an untrusted auxiliary model when its public YouTube audio/visual understanding, public-URL context, or an independent second analysis materially helps Codex answer or verify a request. Use for one or multiple public videos, follow-up questions, cross-checking, or a fresh…

andrewLi1994/gemini-web-bridge · 110 tokens

xrocket-exchange

Use the unofficial xRocket Exchange MCP to inspect spot markets and accounts, trade autonomously inside an operator-configured daily value limit, or prepare explicitly approved transfers and withdrawals. Trigger when a request concerns xRocket markets, account data, deposits, trading, transfers, withdrawals, or its…

nakazanie-ton/myrocket · 87 tokens

apple-mail

Manage Apple Mail on macOS with Gmail-like draft resources, future send scheduling, and explicit sending identities. Use for searching or reading mail, listing sender aliases, selecting a From identity, creating/reviewing/updating/sending/scheduling/deleting drafts, replying, forwarding, organizing messages, or…

yu2001-s/apple-mail · 67 tokens