remember

A way to record the reasoning behind important technical decisions, lessons, and lasting project constraints.

In plain words
What is it for?
Use it to document why one viable option was chosen, what failed, or which constraint future maintainers must understand.
Why use it?
Code and commit history often show what changed but not why, making old decisions costly to reconstruct.

Skill for Claude CodeCodex

Part of the chamnan plugin — 8 skills, 2 agents, 4 hooks shipped together

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.

agentmods
npx agentmods add skills/arcticfox2029/chamnan/remember
Any agent
npx skills add ArcticFox2029/chamnan --skill remember
Clone the repo
git clone --depth 1 https://github.com/ArcticFox2029/chamnan

Made for: Claude Code, Codex.

Or install chamnan, the plugin that ships this one along with the rest of its 8 skills, 2 agents, 4 hooks.

Per session 40 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 1,255 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
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 $0.00040 $0.01255
Opus 5 $0.00020 $0.00628
Sonnet 5 $0.00008 $0.00251
Haiku 4.5 $0.00004 $0.00126

Measured 2d ago against content hash 997c8ba14524, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

remember 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.

skills/remember/SKILL.md · 117 lines

How it starts

The opening of the file, as written. The whole thing — 117 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Write down why

The code says what. The git history says when. Neither says why, and why is the expensive part to reconstruct — usually at the moment somebody is about to undo it.

When to write an entry

Write one when the answer to "why is it like this?" took real work to arrive at, and would take real work to arrive at again.

Good reasons to write:

  • a choice was made between two viable options and the loser is not obvious from the code
  • something was tried, failed, and the failure is not visible in the repository
  • a constraint exists that a newcomer would break without knowing

Do not write an entry when:

  • the code already says it — a well-named function does not need a memory entry
  • the git history already says it — a good commit message is the right place for "what changed"
  • it is a task, not knowledge — that is STATE.md or a session record
  • it is a procedure — that is /chamnan:capture

This is not a conversation log. Never paste a transcript. Write the conclusion.

Pick a category

The three are used differently, which is why they are separate:

decisions/ A choice and its reasoning "Postgres over SQLite — two writers, and SQLite's locking showed up under load."
lessons/ Something that cost time once "Editing src/ while the app runs throws an AttributeError. It is a hot-reload artifact, not a bug — restart clears it."
rules/ A standing constraint "Never add a Cloud fallback for embeddings — a different model means an incompatible vector space."

Rules are injected into every session. Decisions and lessons contribute their title, and the agent reads the file when the title looks relevant. So put a constraint in rules/ only if it should be in front of the agent before it starts — otherwise it is a decision, and the difference matters to what every session costs.

Where it goes

.chamnan/memory/<category>/<short-slug>.md

A short, specific filename: postgres-over-sqlite.md, not database.md. The filename is what appears in the session listing beside the title, so it should be recognisable on its own.

Read the full file on GitHub · 117 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. 2d ago First seen · 117 lines · 40 tokens per session scan A 997c8ba14524

Subscribe to this mod's changes

remember is a skill published in the GitHub repository ArcticFox2029/chamnan (4 stars, last pushed 4d ago), licensed MIT. It adds 40 tokens to every session and 1,255 once invoked, about $0.0002 per session on Opus 5. 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.

Related

Other skills, from other repositories