penfield

penfield is a skill for Claude Code from penfieldlabs/claude-penfield. It costs 77 tokens per session (2,297 once invoked), scanned B, original, MIT.

A persistent memory system for saving and finding project context, decisions, preferences, and related knowledge across sessions.

In plain words
What is it for?
Remembering past work, recalling decisions, connecting related knowledge, saving session handoffs, and managing artifacts.
Why use it?
It prevents important information from being lost when a session ends and helps recover earlier context when needed.

Skill for Claude Code

Written for Claude Code: shipped in a Claude Code plugin. Also seen: reads .claude/ paths; mentions Claude Code.

Part of the penfield plugin — 1 skill, 1 command, 2 hooks, 1 MCP server shipped together

Good fit Remembering past work, recalling decisions, connecting related knowledge, saving session handoffs, and managing artifacts.

Compare 6 skills from other repositories ↓
Install with agentmods
npx agentmods add skills/penfieldlabs/claude-penfield/penfield
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.

Any agent
npx skills add penfieldlabs/claude-penfield --skill penfield
Clone the repo
git clone --depth 1 https://github.com/penfieldlabs/claude-penfield

Made for: Claude Code.

Or install penfield, the plugin that ships this one along with the rest of its 1 skill, 1 command, 2 hooks, 1 MCP server.

Wrote this? Show the measurements

A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.

agentmods badge for penfield

README.md
[![agentmods](https://agentmods.dev/badge/skills/penfieldlabs/claude-penfield/penfield/github.svg)](https://agentmods.dev/skills/penfieldlabs/claude-penfield/penfield)
Your own site
<a href="https://agentmods.dev/skills/penfieldlabs/claude-penfield/penfield"><img src="https://agentmods.dev/badge/skills/penfieldlabs/claude-penfield/penfield/github.svg" alt="Measured on agentmods" height="20"></a>

Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.

agentmods 80×15 button for penfield

Your own site · 80×15
<a href="https://agentmods.dev/skills/penfieldlabs/claude-penfield/penfield"><img src="https://agentmods.dev/badge/skills/penfieldlabs/claude-penfield/penfield.svg" alt="Reviewed on agentmods" width="80" height="20"></a>
Per session 77 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 2,297 The whole file, excluding the scripts and references it only reads on demand.
Security scan B 1 finding. A grade says what 26 rules found in the file — not that it is safe.
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 5.1 $0.00077 $0.02297
Opus 5 $0.00039 $0.01149
Sonnet 5 $0.00015 $0.00459
Haiku 4.5 $0.00008 $0.00230

Measured 10d ago against content hash 437ffb6b27b0, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-10, from the pricing page.

Security

Grade B, and why

penfield scanned grade B with 1 finding 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 10d 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.

Reads agent configuration directoriesmediumAgent snooping

.claude/, .codex/, .gemini/ hold keys, settings and other credentials a mod has no legitimate need for.

cat ~/.claude/projects/*/*.jsonl | jq -r 'select(.message.content) | .message.content' | grep -i "keyword"
plugin/skills/penfield/SKILL.md · 244 lines

How it starts

The opening of the file, as written. The whole thing — 244 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Penfield Memory System

You have access to Penfield, a persistent memory system with hybrid search (BM25 + vector + graph), knowledge graphs, and context management. Memory persists across all sessions and platforms.

Session Lifecycle

Starting a Session

  1. Call awaken() to load personality configuration and user preferences
  2. Call reflect({ time_window: "recent" }) to see what you've been working on lately
  3. Let this context shape your first response — don't greet cold

During a Session

  • Before answering questions about past work: Call recall() first — check what you already know
  • When important things happen: Store decisions, corrections, breakthroughs, and user preferences
  • After storing: Connect related memories using connect() — no orphan memories
  • When the user says "remember this": Always store immediately

Ending a Session

  • If substantive work was done, call save_context() with a detailed cognitive handoff:
    • What you were investigating
    • Key discoveries and decisions
    • Current hypotheses and open questions
    • Suggested next steps
    • References to specific memory IDs

Memory Types (11)

Memory type is auto-detected from content — you don't pass it as a parameter. Write descriptively so the system classifies correctly. Here's what each type represents:

Core Types (8)

Type When to Use
fact Verified, durable information — preferences, specs, organizational data, stable reference material
insight Patterns, realizations, conclusions drawn from observations — discovered patterns, analysis findings, non-obvious connections
correction Fixes to prior understanding — misconceptions corrected, outdated assumptions updated, "we thought X but actually Y"
conversation Session summaries, notable exchanges — important discussions, decision-making contexts, how conclusions were reached
reference Source material, citations, external documentation — RFCs, spec links, documentation URLs, quoted external sources
task Work items, action items, todos — pending work, follow-ups, investigation needs, deferred decisions
strategy Approaches, methods, plans — problem-solving approaches, workflow preferences, codebase mental models, repeatable processes
checkpoint Milestone states, progress markers — project progress snapshots, phase completions, handoff points

Read the full file on GitHub · 244 lines

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. 10d ago First seen · 244 lines · 77 tokens per session scan B 437ffb6b27b0

Subscribe to this mod's changes

penfield is a skill published in the GitHub repository penfieldlabs/claude-penfield (2 stars, last pushed 7mo ago), licensed MIT. It adds 77 tokens to every session and 2,297 once invoked, about $0.0004 per session on Opus 5. A static security scan graded it B with 1 finding (reads agent configuration directories). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.