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 skills add ralfyishere/rules-with-receipts --skill memory-hygienegit clone --depth 1 https://github.com/ralfyishere/rules-with-receiptsWrote 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.
[](https://agentmods.dev/skills/ralfyishere/rules-with-receipts/memory-hygiene)<a href="https://agentmods.dev/skills/ralfyishere/rules-with-receipts/memory-hygiene"><img src="https://agentmods.dev/badge/skills/ralfyishere/rules-with-receipts/memory-hygiene/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.
<a href="https://agentmods.dev/skills/ralfyishere/rules-with-receipts/memory-hygiene"><img src="https://agentmods.dev/badge/skills/ralfyishere/rules-with-receipts/memory-hygiene.svg" alt="Reviewed on agentmods" width="80" height="20"></a>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.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00097 | $0.01462 |
| Opus 5.5 | $0.00039 | $0.00585 |
| Sonnet 5.5 | $0.00019 | $0.00292 |
| Haiku 4.5 | $0.00010 | $0.00146 |
Grade A, and why
Memory Hygiene 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 1mo 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.
How it starts
The opening of the file, as written. The whole thing — 76 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Memory Hygiene
Purpose
In a long session, your context is a mix of load-bearing facts, expired snapshots, superseded decisions, and things you inferred once and never checked. Treating that mix as uniformly true causes a distinctive class of error: acting confidently on information that was correct. The discipline: know which class each remembered item belongs to, and re-verify anything load-bearing that the world could have changed since you learned it.
When to use this skill
- Long sessions, resumed sessions, and any session after context compaction/summarization — the summary preserved conclusions but dropped the evidence behind them.
- Before acting on something you recall from "earlier": a file's contents, a decision, a constraint, a number.
- When the user says something that contradicts your memory — the user is more current than your context.
- When deciding what to write to persistent memory files, and when reading memories back in a later session.
When NOT to use this skill
- Short sessions where everything relevant is a few messages up and nothing has changed it.
- Don't ritualize: re-verifying a stable user preference every turn is friction, not hygiene. The test is could it have changed, and is it load-bearing?
Operating procedure
Step 1 — Classify before relying. When reaching for a remembered item, tag it:
| Class | Examples | Reliability rule |
|---|---|---|
| Durable fact | The project is in Python; the user's role; decisions explicitly confirmed | Trust; re-verify only on contradiction |
| User preference | Tone, format, "always show diffs", risk appetite | Trust until contradicted; newest statement wins |
| Snapshot | File contents you read, test results, data values, what a doc said | Expires the moment anything could have changed it — incl. your own edits. Re-observe before load-bearing use (live-state-truth) |
| Working assumption | Things inferred to proceed, never confirmed | Still an assumption no matter how long ago you made it; label or verify |
| Superseded | Plans revised mid-session, requirements the user changed | Actively dangerous — the old version is still in your context, fluent and available. Latest statement wins |
| Task-local detail | Variable names from a finished task, one-off constraints | Let it go; do not generalize to new tasks |
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.
- 1mo ago First seen · 76 lines · 97 tokens per session scan A 9448237684c9
Memory Hygiene is a skill published in the GitHub repository ralfyishere/rules-with-receipts (2 stars, last pushed 3mo ago), licensed MIT. It adds 97 tokens to every session and 1,462 once invoked, about $0.0004 per session on Opus 5.5. 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.
Other skills, from other repositories
cross-task-learner
Enable agent loops to learn from similar past tasks and share patterns across loops.
ralph-memory
Manage Al semantic memory entries — list, query, and clear lessons learned across loop iterations.
harness-memory-maintenance
Audit and maintain harness memory through reviewed aiwg memory plans, with owner receipts, verification and recovery.
distill-feedback
A process for turning corrections you give an agent into lasting working rules. It reviews saved conversations and asks for approval before changing those rules.
session-explore
Investigate past AI session activity with cited catalog search, timelines, tool analytics, and bounded comparisons across providers.
reflection-injection
Inject relevant past reflections into agent context at iteration start so agents learn from prior mistakes without repeating them.