Install

brew tap AssafWoo/pandafilter
brew install pandafilter

Linux / any platform:

curl -fsSL https://raw.githubusercontent.com/AssafWoo/homebrew-pandafilter/main/install.sh | bash

First run: PandaFilter downloads the BERT model (~90 MB, all-MiniLM-L6-v2) from HuggingFace and caches it at ~/.cache/huggingface/. Subsequent runs are instant.

Then wire it in — one command installs for every AI agent you have:

panda init --agent all

Auto-detects Claude Code, Cursor, Gemini CLI, Codex, Windsurf, Cline, OpenClaw, and VS Code Copilot. Skips anything that isn't installed. Or target one specifically:

panda init                          # Claude Code (default)
panda init --agent cursor           # Cursor
panda init --agent gemini           # Gemini CLI
panda init --agent codex            # Codex (CLI + VS Code extension)
panda init --agent windsurf         # Windsurf
panda init --agent cline            # Cline
panda init --agent openclaw         # OpenClaw
panda init --agent copilot          # VS Code Copilot

---

or: cargo uninstall panda

rm -rf ~/.local/share/panda # analytics + sessions rm -rf ~/.cache/huggingface/hub/models--sentence-transformers--all-MiniLM-L6-v2 ```

</details>

---

now active automatically — no config change needed.

[focus] enabled = false # disabled by default — enable with panda focus --enable after testing min_files = 25 # skip for repos smaller than this min_lines = 2000 # skip for repos with fewer source lines

[commands.git] patterns = [ { regex = "^(Counting|Compressing|Receiving|Resolving) objects:.*", action = "Remove" }, ]

[commands.cargo] patterns = [ { regex = "^\\s+Compiling \\S+ v[\\d.]+", action = "Collapse" }, { regex = "^\\s+Downloaded \\S+ v[\\d.]+", action = "Remove" }, ]

Pattern actions: `Remove`, `Collapse`, `ReplaceWith = "text"`, `TruncateLinesAt = N`, `HeadLines = N`, `TailLines = N`, `MatchOutput = "msg"`, `OnEmpty = "msg"`.

Pricing uses `cost_per_million_tokens` from `panda.toml` if set, otherwise `ANTHROPIC_MODEL` env var (Opus 4.6: $15, Sonnet 4.6: $3, Haiku 4.5: $0.80), otherwise $3.00.

</details>

<details>
<summary><strong>User-defined filters</strong></summary>

Place `filters.toml` at `.panda/filters.toml` (project-local) or `~/.config/panda/filters.toml` (global). Project-local overrides global for the same key. Runs before any built-in handler.

[commands.myapp.match_output] pattern = "Server started" message = "ok — server ready" unless_pattern = "error"

</details>

<details>
<summary><strong>Session intelligence</strong></summary>

State tracked via `PANDA_SESSION_ID=$PPID`, stored at `~/.local/share/panda/sessions/<id>.json`.

- **Result cache** — post-pipeline bytes frozen per input hash; returned identically on repeat calls to prevent prompt cache busts.
- **Semantic delta** — repeated commands emit only new/changed lines: `[Δ from turn N: +M new, K repeated — ~T tokens saved]`.
- **Cross-turn dedup** — identical outputs (cosine > 0.92) collapse to `[same output as turn 4 (3m ago) — 1.2k tokens saved]`.
- **Elastic context** — pipeline pressure scales with session size. At >80% pressure: `[⚠ context near full — run panda compress --scan-session]`.
- **Intent-aware query** — reads the agent's last message from the live session JSONL and uses it as the BERT query.
- **File delta re-reads** *(v1.3.0)* — re-reading a changed file sends a unified diff instead of the full content. Unchanged re-reads send a structural digest (function/class signatures). Both save 60–95% of re-read tokens automatically.
- **Compaction digest** *(v1.3.0, Claude Code only)* — before Claude auto-compacts, PandaFilter serializes edited files, error signatures, and top commands to `~/.local/share/panda/compacts/`. On the next session start, the digest is injected into context so the agent resumes oriented.

</details>

<details>
<summary><strong>Supported agents (7)</strong></summary>

All agents share the same binary and filtering pipeline. `panda init --agent all` installs for everything detected on your machine in one shot.

| Agent | Install | Config |
|-------|---------|--------|
| Claude Code | `panda init` | `~/.claude/settings.json` |
| Cursor | `panda init --agent cursor` | `~/.cursor/hooks.json` |
| Gemini CLI | `panda init --agent gemini` | `~/.gemini/settings.json` |
| Codex (CLI + VS Code) | `panda init --agent codex` | `~/.codex/hooks.json` |
| Windsurf | `panda init --agent windsurf` | `~/.codeium/windsurf/hooks.json` |
| Cline | `panda init --agent cline` | `.clinerules` (project dir) |
| VS Code Copilot | `panda init --agent copilot` | `.github/hooks/` (project dir) |

**Hook-based agents** (Claude Code, Cursor, Gemini, Codex, Windsurf) intercept every command before and after execution via the agent's native hook system.

**Rules-based agents** (Cline, Copilot) inject `panda run <cmd>` directives into the agent's context file, relying on the model to follow them.

**PreToolUse:** known handler → rewrites to `panda run <cmd>`; unknown → no-op; already wrapped → no double-wrap; compound commands → each segment rewritten independently.

**PostToolUse:** Bash → full pipeline; Read → BERT + session dedup; Glob → grouped by directory; Grep → compact paths.

**UserPromptSubmit:** Context Focusing module → queries file graph → injects guidance (recommended + excluded files).

**Hook integrity:** `panda init` writes SHA-256 baselines (chmod 0o444). PandaFilter verifies at every invocation and exits 1 with a warning if tampered. `panda verify` checks all installed agents.

</details>

<details>
<summary><strong>Crate overview</strong></summary>

</details>

<details>
<summary><strong>Uninstall</strong></summary>

brew uninstall pandafilter && brew untap AssafWoo/pandafilter # Homebrew

FAQ

Does PandaFilter change what the agent can see? It removes noise — build progress, passing test lines, module download logs. Errors, file paths, and results are always kept.

What if I don't want a specific command filtered? Add a rule to .panda/filters.toml to customize or override any handler. See the User-defined filters section. You can also use panda proxy <cmd> to run a command raw with no filtering.

What about commands PandaFilter doesn't know? Output passes through unchanged. PandaFilter never silently drops output from unknown commands.

How do I verify it's working? Run panda gain after a session. To see exactly what the agent received from a specific command: panda run git log --oneline -20.

Does PandaFilter send any data outside my machine? No. All processing is fully local. BERT runs on-device.

What is Context Focusing? An opt-in feature that tells the agent which files are relevant for the current prompt, preventing it from reading unrelated files. Enable with panda focus --enable after running panda doctor to confirm the index is ready.

---