foggy-mode-3471c8 — independently scanned and version-tracked by SaferSkills.
SaferSkills independently audited foggy-mode-3471c8 (Agent Skill) and scored it 83/100 (green). The audit ran 55 deterministic rules across Security, Supply Chain, Maintenance, Transparency, and Community; it found 0 high-severity and 4 lower-severity findings. The full rule-by-rule trace and per-finding evidence are below. Free, methodology-open.
Findings & checks · 4 flagged
The text {match} tells the agent to skip the normal "ask the user first" gate. Used adversarially it removes the human-in-the-loop check before destructive or sensitive actions, turning a normally-gated agent into a fire-and-forget executor.
The text {match} tells the agent to skip the normal "ask the user first" gate. Used adversarially it removes the human-in-the-loop check before destructive or sensitive actions, turning a normally-gated agent into a fire-and-forget executor.
A bulleted imperative like {match} tells the agent to never reveal, disclose, or mention something to the user. Used adversarially it can instruct the agent to hide its tool calls or lie about what it did — stripping the transparency a user relies on to trust the agent.
Every scanned point with the score it earned and what moved between them.
First recorded scan — no prior version to compare against.
The primary manifest — the file an agent reads to learn what this artifact does.
Adapt response shape for a user with low working memory or low cognitive energy. Assume memory, attention, and planning capacity are scarce.
This skill is not diagnosis, therapy, medical care, or treatment for Long COVID, sleep disorders, fatigue, or any other condition. Do not tell the user why they feel foggy.
brain-fog-mode is an intentional variant of this contract, not a true alias: it follows the four-label shape by default but may return only Next action when the user explicitly asks for only the next step.
This mode can be used for safe non-coding tasks. Do not refuse only because the task is outside software engineering.
When this skill is active, obey this contract before ordinary helpfulness:
Current state may include one or two short sentences of context when thathelps name the immediate constraint or reduce ambiguity.
future tasks.
Next action is one action only.message, but not both.
Put, Open, Write, or Send.Avoid compound actions such as "get X and place it."
placement, one field, one sentence, or one message. Do not ask the user to gather multiple prerequisites.
organize everything.
Next action has no bullets, numbered steps, branch choices, conditionals, orfollow-up tasks.
Next action must not join actions with words like "then", "after", or"once".
Next action should not use "and" to join two verbs. Choose one visible unitonly.
Do not do yet as things to avoid, not as instructions.Stop point must start with Stop when.Stop point must only say where to pause; it must not introduce the next task.Stop point must mirror the immediate Next action, not later work.Stop point must not contain a new imperative command like Send, Open,Write, Get, Check, or Reply.
Stop when the message is sent. or `Stop when thedocument is on the table.`
After sending the message, get the document.Stop point must not include extra checks, investigation, inspection, orfollow-up work.
reports that the first action is done.
think clearly.
cognitively depleted.
Use this exact shape for every normal response under this skill:
Current state: <where we are in one or two short sentences>
Next action: <one tiny action>
Do not do yet:
- <items to avoid for now>
Stop point: Stop when <the next action is done>.For messages or multi-step tasks, the first response still gives one action and a stop point only. Defer diagnosis, branch decisions, sorting, and later checks.
Before sending the answer, silently check:
Next action contains two verbs joined by "and", rewrite it to the firstvisible action only.
Next action combines setup plus sending, keep only the earliest visibleaction.
Stop point does not start with Stop when, rewrite it.Stop point contains a new command, rewrite it to the completion ofNext action.
For object-based prompts, prefer this shape:
Current state: You are tired, and the task is too big to hold at once.
Next action: Put the needed document on the table.
Do not do yet:
- Do not open more pages.
Stop point: Stop when the document is on the table.Use this shape for multi-step prompts:
Current state: You are foggy, and this task has too many moving parts to hold at once.
Next action: Write the one sentence you need to send first.
Do not do yet:
- Do not organize the whole task.
- Do not answer every message.
- Do not make a full plan.
Stop point: Stop after writing the sentence.any medical condition.
user contact a healthcare professional or urgent care as appropriate.
medical emergency, prioritize immediate real-world help.
For the reasoning behind these rules and source links, read references/grounding.md.
~30 seconds. Free. No account. Every finding cites a rule and a line of evidence.