decorated-pi
decorated-pi is a practical enhancement pack for pi coding agent — token-efficient workflow, cache-friendly design, and smarter tools.
6.2k listings for pi · 0 reviewed by us
323 results · page 2 of 7
decorated-pi is a practical enhancement pack for pi coding agent — token-efficient workflow, cache-friendly design, and smarter tools.
A pi extension that prompts before dangerous bash commands and protected file writes.
Evidence-first vulnerability research skills and CLI for Pi, Codex, and Claude Code
Hard-block dangerous agent bash commands and secret path access
Append-only project-local learning ledger for Pi with two separate trust gates: manual human approval, and a bounded auto-safe machine gate that is on by default in Pi-trusted projects. Selective fork of Matt Devy's pi-continuous-learning.
A pi extension that prompts before session changes when the current git repo has uncommitted changes.
High-risk-only approval hook with LLM risk analysis, behavior detection, protected-path interception, and decision memory for oh-my-pi (OMP) and pi-agent.
Pi extension enforcing no-implementation during wayfinder (fog mode), with a parallel four-axis code review.
Context engine for Pi — zero-LLM compaction, session continuity, sandbox execution, and tool display optimization
Track LLM spending per task and pause when a configurable threshold is reached — continue, refine the prompt, or stop. Pi extension by agentoom.com.
Permission enforcement extension for pi — gates tool calls, bash commands (via arbiter executable), and file paths
Pi extension that detects and renames non-ASCII filenames before fff-core can panic on UTF-8 byte boundaries.
Official ryk Pi integration for fail-closed bash and file protection.
Research-driven planning, debugging, and audit skills for Codex, Pi, and Claude Code
Model-backed boundary approval broker for Pi
Codex-style automatic approval reviews for @gotgenes/pi-permission-system
Small Pi safety extension for cwd access, protected paths, runtime config confirmation, and common destructive bash commands.
A read-only two-stage code review extension for pi
Pi extension that reduces Agent Skill catalog tokens with policy-based skill retrieval and on-demand skill loading.
Context-aware Bash permissions for Pi with automated guardian review.
Pi coding agent extension that runs all tool calls inside a remote Upstash Box sandbox while the agent runs locally
Secret-safe workflow reminder that steers Pi agents to use nopeek instead of exposing .env values
Agents should invoke this skill for broad multi-claim research projects needing planning, parallel investigation, source merging, gap closure, citation audit, and final synthesis when narrower research skills are insufficient.
sudo for pi — agent asks, user types their password into a masked TUI input, password never reaches the model
High-performance secret firewall for pi — secrets usable as shell vars, never visible to models
Four approval modes + command-level security restrictions for the Pi coding agent. Act / Auto / Ask / Plan mode switching, rule engine, subagent auto-approval.
Blocking, user-approved sudo tasks in a dedicated Herdr pane
Code-review workflow skills for pi sessions — code-review (AI-powered review with severity labels via the CodeRabbit CLI) and autofix (safely review and apply CodeRabbit PR review-thread feedback with per-change approval). Pure-skill package, no extension
A pi extension that runs an isolated repo workflow audit and returns only the final report to the main session.
A pi extension that scans locally installed pi packages for vulnerabilities and supply-chain risk.
Pi extension that guards agent network and website access with global and project allow lists.
macOS Seatbelt sandbox extension package for pi bash and file-tool access controls.
Pi package scaffold for pi-prompt-guard
Native Pi extension: give a Pi coding agent a real, routable Whisper IPv6 /128 identity - keyless verify/RDAP; the full control plane, /128 egress, and the whisper.security graph (Cypher + 29 recipes) with a key.
Interactive cybersecurity news briefing for the Pi coding agent with live multi-source headlines and deep-dive story prompts
Sandbox write/edit/bash to the project dir — prompt / deny / allow modes, configurable deny-with-redirect rules, per-session allow-deny memory
Run Pi coding tools in a Celesto computer
Pi extension that detects and prevents LLM loops in tool calls and reasoning
A Claude safety-classifier refusal shouldn't kill your turn — rescue, retarget the server-side fallback chain, and see what's tripping.
Global tool-output redaction for Pi — redacts secrets, certificates, PII, and connection strings across all tools before they reach the model.
Host package for managing one shared Pi widget slot across multiple providers with preset policies and a built-in demo provider.
Tab 键切换计划/执行模式
TUI secret prompt for pi — agent asks, user types in a masked input, value never reaches the model
Pi extension: bottom-panel ask_user + bash permission gate (Allow / Allow permanently / Deny with reason) writing ~/.pi/agent/permission.json
Self-contained Pi package for skill lifecycle management: memory, skill-bank audit, evaluation, creation, and refinement tools.
Review Excel files for spreadsheet errors — mechanical checks (openpyxl) + Panko–Halverson taxonomy reasoning
Pi extension that protects sensitive files, blocks secret writes, and optionally redacts protected read output.
Run Pi coding-agent tool calls inside Docker sbx sandboxes