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/pivanov/agents-wire/pigit clone --depth 1 https://github.com/pivanov/agents-wireWhat 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 | $0.00451 |
| Opus 5 | $0.00000 | $0.00226 |
| Sonnet 5 | $0.00000 | $0.00090 |
| Haiku 4.5 | $0.00000 | $0.00045 |
Grade A, and why
pi 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
Pi
Agent ID: pi
Pi coding agent by Mario Zechner. Pi v0.73 does not implement the Agent Client Protocol - its --mode rpc uses a Pi-specific JSON dialect that is not ACP-compatible. The catalog entry is preserved so detect still surfaces Pi as installed, but live model resolution is skipped and session prompts will fail. Use Pi directly via the pi CLI until ACP support lands.
::: warning Non-ACP agent
acpCompatible is set to false for Pi. resolveModels() skips the session probe for Pi and shows the static "Default" placeholder. Do not rely on agents-wire to drive Pi until a future release adds ACP support.
:::
Install
npm install -g @mariozechner/pi-coding-agent
export PI_API_KEY="your-api-key"
Quick Start
import { agents } from "@pivanov/agents-wire";
const result = await agents.ask(
"pi",
"Explain this codebase",
{
permission: "auto-allow",
},
);
console.log(result.text);
Model Selection
Pi is a single-model product. options.model and options.effort are not exposed and
have no effect.
Capabilities
| Feature | Supported |
|---|---|
ask / stream / session |
❌ (fails until Pi adds ACP support) |
askJson |
❌ |
| MCP stdio | ❌ |
| MCP http/sse | ❌ |
Session listing (listSessions) |
❌ |
| Tool call interception | ❌ |
Cost Tracking
Pi is subscription-based and does not report per-turn costUsd. cost.snapshot.totalUsd stays at 0, so maxCostUsd does not trigger. Monitor usage with cost.turnCount.
Gotchas
PI_API_KEYmust be set before use.- Session listing is not supported;
listSessions()throwsCapabilityNotSupportedError. - No per-turn cost data - Pi operates on a subscription model.
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 · 59 lines · 0 tokens per session scan A 38557e934216
pi is an agent published in the GitHub repository pivanov/agents-wire (5 stars, last pushed 3mo ago), licensed MIT. It costs nothing until one of its globs matches a file; then it loads 451 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 2026-08-31.
Other agents, from other repositories
codex-execute
Implement a SPECIFIC, well-defined plan or plan-step with Codex. Use ONLY when the user has a written plan, plan-step, named files, or acceptance criteria to execute ("execute this via codex", "implement this plan", "/codex:execute", or a --plan file is in context). Do NOT use when the user is stuck/exploring (use…
cursor-delegate
Delegate a SPECIFIC, well-defined implementation task or plan step to Cursor in agent mode on the Auto model. Cursor is the fast lane for mechanical writing — long file writes (200+ lines), pattern-following across many files, bulk refactors. Supports autonomous multi-step runs via --until-done. Pair with…
cursor-research
Delegate read-only EXTERNAL web/documentation research to Cursor. Use when the user wants current library/API/docs research, "what's the current way to…", or to compare approaches — and you'd rather not spend Claude's context on web reading. Cursor runs read-only (ask mode) with web search + fetch and returns findings…
opencode-research
Delegate read-only EXTERNAL web/documentation research to OpenCode. Use when the user wants current library/API/docs research, "what's the current way to…", or to compare approaches — and you'd rather not spend Claude's context on web reading. OpenCode runs read-only (ask mode) with web search + fetch and returns…
codex-rescue
Hand an OPEN-ENDED or stuck problem to Codex for independent investigation. Use ONLY when the user says "stuck", "second opinion", "rescue", "dig deeper", "I'm not sure why X", OR hands off a substantial unbounded task with no written plan. Do NOT use when a plan or plan-step file is in context (use codex-execute) or…
cursor-explore
Delegate quick read-only CODEBASE exploration to Cursor. Use to answer questions about the current repository ("where is X handled?", "how does Y flow?", "summarize the adapter layer") without spending Claude's tokens reading files. Cursor runs read-only (ask mode) using semantic search, grep, and its Explore…