Borrowing it
Nothing to install: this file belongs to neoneye/agent-memory-atlas. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.
curl -O https://raw.githubusercontent.com/neoneye/agent-memory-atlas/main/.agents/skills/remove-meta-narrative/SKILL.mdgit clone --depth 1 https://github.com/neoneye/agent-memory-atlasWrote 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/neoneye/agent-memory-atlas/remove-meta-narrative)<a href="https://agentmods.dev/skills/neoneye/agent-memory-atlas/remove-meta-narrative"><img src="https://agentmods.dev/badge/skills/neoneye/agent-memory-atlas/remove-meta-narrative/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/neoneye/agent-memory-atlas/remove-meta-narrative"><img src="https://agentmods.dev/badge/skills/neoneye/agent-memory-atlas/remove-meta-narrative.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector pass
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 | $0.00078 | $0.03686 |
| Opus 5 | $0.00039 | $0.01843 |
| Sonnet 5 | $0.00016 | $0.00737 |
| Haiku 4.5 | $0.00008 | $0.00369 |
Grade A, and why
remove-meta-narrative 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 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.
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 — 236 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Remove Meta-Narrative
A published page states what is true. A ## History entry states what
changed. When the second leaks into the first, the reader has to reconstruct
the project's sequence of positions before they can find out what it currently
thinks — and they were only ever asking the second question.
This has been corrected in this repository at least six times, including one sweep across 273 files, and it recurs because the rule is easy to agree with and the failure does not look like a violation while you are committing it.
The shape of the failure
Almost always: appending instead of replacing. A new fact arrives. The natural edit is to write a paragraph stating it and leave the paragraph it supersedes in place. Both are now on the page, the newer one hedged into a transition — "the argument has started to fail", "this is no longer none" — and the page argues with itself. Nothing was deleted, so nothing felt wrong.
Five forms, in increasing order of how well they hide:
| Form | Looks like | Caught by |
|---|---|---|
| Freshness stamp | "Corpus last extended 6 August 2026", "last updated ", "current as of " | grep 5 |
| Explicit | "This report previously said", "re-read on ", "the atlas had missed" | grep 1 |
| Self-report | "this page said '164 open-source systems'", "the count was N until 7 August", "the sentence was false in the way that matters" | grep 4 |
| Adverbial | "there are now two paths", "still inverts", "no longer none", "has since been read" | grep 2 |
| Structural | a paragraph superseded by the next one, both retained; a bullet holding two positions | nothing automatic — you have to read it |
The freshness-stamp form is noise even when nothing is wrong. A line dating
when the page, the corpus or a section was last extended, updated, refreshed or
reviewed stamps the atlas's own bookkeeping onto a page whose job is the current
state of memory systems — and it is stale within days of the next commit while
telling the reader nothing they came to learn. Unlike the other four, it is not
narrating a changed position; it is a self-dating currency stamp, and it reads as
cringe precisely because the reader can see it has already rotted. The corpus
counts are live — generated and build-checked — so the page is as current as its
last build; announcing that with a hardcoded date only decays. If when a reading
happened genuinely matters, it belongs in a report's dated ## History, never
stamped on the compare page. Cut the stamp; do not "update" it, because the next
edit re-stales it.
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.
- 10d ago First seen · 236 lines · 78 tokens per session scan A 1bff75ec4bcc
remove-meta-narrative is a skill published in the GitHub repository neoneye/agent-memory-atlas (64 stars, last pushed today), licensed MIT. It adds 78 tokens to every session and 3,686 once invoked, about $0.0004 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 2026-08-30.
Other skills, from other repositories
thoughtline-memory
Persistent, project-aware memory for AI coding agents. Use thoughtline whenever you make a decision, fix a bug, learn something non-obvious, or need to recall prior work.
recall
Recall this repository's OwnMem local memory before changing code, and keep it healthy. Use when a repository contains .ownmem/, when past debugging lessons could apply ("have we hit this before", "why is it done this way"), or when the user mentions ownmem, project memory, or recalling across sessions.
Effective Memory
The essential habits for an AI agent with memory — session bookends, learning triggers, verification, safety, and the operational discipline that turns raw recall into compounding intelligence. Pinned, always-injected.
init
Install or update OwnMem in the current repository. Use when the user asks to set up OwnMem, add local project memory for coding agents, or refresh an existing OwnMem installation after a version bump.
honcho-integration
Integrate Honcho memory into existing Python or TypeScript codebases. Use when adding Honcho SDK, setting up peers, configuring sessions, and accessing Honcho's representation.
honcho-memory
Concepts and strategy for using a connected Honcho as persistent memory of the user — the recall/record loop and session and peer design. Start here to understand how Honcho memory works, then connect — via a first-class integration for your environment if one exists (preferred), or raw MCP tools (covered here) or the…