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 agents/marshalleriksen-neura/ragcode/cli-adaptersgit clone --depth 1 https://github.com/MarshallEriksen-Neura/ragcodeWhat 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 | $0.01439 |
| Opus 5 | $0.00000 | $0.00720 |
| Sonnet 5 | $0.00000 | $0.00288 |
| Haiku 4.5 | $0.00000 | $0.00144 |
Grade A, and why
cli-adapters 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.
How it starts
The opening of the file, as written. The whole thing — 122 lines — stays where its author put it; the contents beside it link to each section on GitHub.
RagCode Agent CLI Adapters
RagCode orchestrates AI agents by spawning their CLIs. Vendor CLI syntax
drifts, so built-in command templates are presets, not hard-coded
protocol truth. Anything can be overridden per-repo via
.ragcode/agents/agents.json.
Built-in presets
| Agent | Command | Default args | supportsWrite | connectorKind |
|---|---|---|---|---|
| codex | codex |
--ask-for-approval never |
true | codex-app-server (Tier 1) |
| claude | claude |
(none) | true | claude-stream-json (Tier 2) |
| gemini | gemini |
(none) | false | generic CLI spawn |
| grok | grok |
(none) | true | grok-agent-stdio (Tier 2) |
Grok uses grok agent stdio (ACP JSON-RPC). Probe only checks the subcommand
exists; a real turn requires grok login (OAuth / cached token).
| Concern | Behavior |
|---|---|
| Stream | session/update → agent_message_chunk (coalesced into one harness message) |
| Terminal | _x.ai/session_notification / session/prompt stopReason |
| Resume | Start-time only: start({ nativeSessionId }) → session/load({ sessionId, cwd, mcpServers }). No silent fallback to session/new. Post-start resume() throws. |
| Cancel | ACP session/cancel notification, then process-tree kill (authoritative). Exactly one harness terminal; late stream events dropped. |
| Approval | Strict: session/request_permission → ACP { outcome: { outcome: "selected", optionId } }. With permission_mode=always-approve (common default), Grok auto-approves and only emits pending_interaction / interaction_resolved notifications — no client RPC. |
Capture lab: npx tsx scripts/grok-acp-capture.ts --mode resume|cancel
Notes: docs/agents/captures/grok-acp-control-plane.md
Configuration: .ragcode/agents/agents.json
{
"agents": {
"codex": { "adapter": "cli", "command": "codex", "args": ["--ask-for-approval", "never"], "supportsWrite": true, "supportsJson": true, "supportsStreaming": true, "supportsMcp": true, "defaultTimeoutMs": 600000, "connectorKind": "codex-app-server" },
"claude": { "adapter": "cli", "command": "claude", "args": [], "supportsWrite": true, "supportsJson": true, "supportsStreaming": true, "supportsMcp": true, "defaultTimeoutMs": 600000, "connectorKind": "claude-stream-json" },
"gemini": { "adapter": "cli", "command": "gemini", "args": [], "supportsWrite": false, "supportsJson": true, "supportsStreaming": true, "supportsMcp": false, "defaultTimeoutMs": 300000 },
"grok": { "adapter": "cli", "command": "grok", "args": [], "supportsWrite": true, "supportsJson": true, "supportsStreaming": true, "supportsMcp": true, "defaultTimeoutMs": 600000, "connectorKind": "grok-agent-stdio" }
}
}
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.
- 2d ago First seen · 122 lines · 0 tokens per session scan A fe836b91e573
cli-adapters is an agent published in the GitHub repository MarshallEriksen-Neura/ragcode (11 stars, last pushed 1mo ago), licensed MIT. It costs nothing until one of its globs matches a file; then it loads 1,439 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 agents, from other repositories
sverklo-explore
Drop-in replacement for Claude Code's built-in Explore subagent. Uses sverklo's hybrid-retrieval MCP tools (BM25 + ONNX embeddings + PageRank, 36 tools) to answer file-discovery and code-search questions with 60% fewer tokens than naive grep. Use this when you need to locate definitions, trace references, understand…
reviewer
Independent semantic review of a Cartograph Rust/PostgreSQL diff before merge or release. Read-only; checks correctness, bounds, concurrency, migration safety, agent evidence quality, and gate alignment.
README
Adapted from the TORQ Console harness (E:\TORQ-CONSOLE\.claude\agents). These are TorqClaw-local definitions. The Console harness drives a different product through .torq/launch/ watcher scripts and hardcoded E:\TORQ-CONSOLE\... paths; running those agents against this repo would read specs from and write verdicts…
domain
How the engineering skills should consume this repo's domain documentation when exploring the codebase.
Dashboard Dev
Agent "Dashboard Dev" from shawnhack/exocortex, covering dashboard dev, core responsibilities, key files, coding constraints and escalation.
Memory Engineer
Agent "Memory Engineer" from shawnhack/exocortex, covering memory engineer, core responsibilities, key files, coding constraints and escalation.