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 Filipe-Soares-de-Almeida/MemAI --skill memai-confidencegit clone --depth 1 https://github.com/Filipe-Soares-de-Almeida/MemAIWrote 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/filipe-soares-de-almeida/memai/memai-confidence)<a href="https://agentmods.dev/skills/filipe-soares-de-almeida/memai/memai-confidence"><img src="https://agentmods.dev/badge/skills/filipe-soares-de-almeida/memai/memai-confidence/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/filipe-soares-de-almeida/memai/memai-confidence"><img src="https://agentmods.dev/badge/skills/filipe-soares-de-almeida/memai/memai-confidence.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 | $0.00155 | $0.02604 |
| Opus 5 | $0.00077 | $0.01302 |
| Sonnet 5 | $0.00031 | $0.00521 |
| Haiku 4.5 | $0.00015 | $0.00260 |
Grade A, and why
memai-confidence 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 12d 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 — 190 lines — stays where its author put it; the contents beside it link to each section on GitHub.
memai-confidence — verifying facts, setting confidence
confidence is only worth setting when the claim was checked live. This
skill reads a memory's anchors against today's source and sets the field —
and it is where a memory the store handed back as due gets its recheck
(§5).
The pass proposes: everything below is staged through optimize_stage and
applied by a human in the dashboard.
1. The three values
| Value | When |
|---|---|
confirmed |
A note/reasoning/anti_pattern/diagram whose structural anchors were checked live: the file and line exist, the symbol is there, the table/column exists, the domain resolves, the URL answers. |
contradicted |
The claim is false today — the fact no longer holds. Requires a non-empty verified; normally staged next to a reword or an archive. |
unverified |
Nothing to check it against: ephemeral or point-in-time (a count, a value, "open this month"), external, or a checkpoint still bearing live work. |
A checkpoint stays
unverified. Its content is where the work stood, not a verifiable fact, so there is nothing for a live check to agree with. A checkpoint for an effort that has closed is not a confidence question either — it is retirement, in [[memai-checkpoints]].
Do not let an ephemeral tail contaminate a durable core. A note carrying a durable mechanism plus one dated example (the counts from one run, one queue's depth) is
confirmedon the mechanism. The dated example neither blocks the confirmation nor earns a demotion.
2. How a check is made
- Start from the
anchorsthe scan already extracted. Per memory,optimize_scanpulls the verifiable references out of the full body — URLs, file paths, table/field-style identifiers,SNAKE_CASEconstants — and returns them space-joined. That is the work list, without re-reading every body. It is capped at the top 5: when the verdict needs more than those, open the record withget_memory(uid). - A check is read-only, against the live source the claim points at. Code and config: read the file, grep the symbol. Schema: describe the table. A URL: fetch it. Which tool is allowed to read a database, and which agent a fan-out may be delegated to, are the host project's rules — read them and use what they approve. This skill names no server and no model tier; it only requires that the check touches the real thing and changes nothing.
- Fan the checks out into separate agents, one per cluster (§4). Verifying an anchor costs the reading it takes — files opened, tables described, dead ends — and that reading belongs in an agent's context, not in the pass's own. Each agent returns a verdict per claim plus the evidence for it; the pass keeps the verdicts.
- Ask for a verdict, not a repair. The agent reports what it found
(
file:line, the column, the resolved domain). Fixing the code is not this pass. - Hand over exact paths. Resolve them (
Glob) before dispatching, or every agent spends its first turns finding what the pass already knew.
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.
- 12d ago First seen · 190 lines · 155 tokens per session scan A 1f69d127bd4b
memai-confidence is a skill published in the GitHub repository Filipe-Soares-de-Almeida/MemAI (0 stars, last pushed yesterday), licensed MIT. It adds 155 tokens to every session and 2,604 once invoked, about $0.0008 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.
Other skills, from other repositories
browserwing-executor
Control browser automation through HTTP API. Supports page navigation, element interaction (click, type, select), data extraction, accessibility snapshot analysis, screenshot, JavaScript execution, and batch operations.
dev-browser
Browser automation with persistent page state. Use when users ask to navigate websites, fill forms, take screenshots, extract web data, test web apps, or automate browser workflows. Trigger phrases include "go to [url]", "click on", "fill out the form", "take a screenshot", "scrape", "automate", "test the website"…
ask-user-question
Ask users questions via the UI. Use when you need clarification, user preferences, or confirmation before proceeding. The user CANNOT see CLI output - this tool is the ONLY way to communicate with them.
memos-memory-guide
Use the MemOS Local memory system to search and use the user's past conversations. Use this skill whenever the user refers to past chats, their own preferences or history, or when you need to answer from prior context. When auto-recall returns nothing (long or unclear user query), generate your own short search query…
memmachine-memory
Use when an agent or model needs durable project, user, or session context from MemMachine, needs to save information to MemMachine memory, has requests involving mem-cli, memmachine, or memmachineclient, has insufficient conversation context, or is tempted to search local files for prior context that should come from…
browserwing-admin
Manage and operate BrowserWing — an intelligent browser automation platform. Install dependencies, configure LLM, create/manage/execute automation scripts, use AI-driven exploration to generate scripts, browse the script marketplace, and troubleshoot issues.