Chock.sh Build your plugin

Policies

Scan Hidden Content

Scan Hidden Content: what it does, its label on each agent, what it misses, and examples.

Catalog policy scan-hidden-content, version 0.0.2, labelled warns on Claude Code.

What it does

Warns (observe) when a change adds text a reader cannot see, or a URL that carries data out, to Markdown, HTML, SVG, XML, Word, agent instruction files, docs and templates: instruction-like or long comments, CSS-hidden, white or 1pt text, hidden Word runs, URLs with secret words, long queries, placeholders or encoded parts, remote embeds, camo runs, HTML data URIs. Runs: commit, agent write, turn's end. Waiver: a person's marker comment line above. Friction, not a boundary.

On each agent

AgentLabelWhat it does there
Claude CodeWarnsWarns on an agent's file writes and at turn end
CursorWarnsWarns on an agent's file writes and at turn end
CodexWarnsWarns on an agent's file writes and at turn end. If the hook errors, the action goes through (not witnessed; the vendor's row is only partly live-run)
GitHub CopilotWarnsWarns at turn end only. It does not check a file write when it happens. If the hook errors, the action goes through (not witnessed; the vendor's row is only partly live-run)
DevinWarnsWarns at turn end only. It does not check a file write when it happens. If the hook errors, the action goes through (vendor docs only, not witnessed)
GrokNot gradedNot graded yet for Grok
gooseNot gradedNot graded yet for goose
AntigravityNot gradedNot graded yet for Antigravity
OpenHandsNot gradedNot graded yet for OpenHands
Kimi CodeNot gradedNot graded yet for Kimi Code
Factory DroidNot gradedNot graded yet for Factory Droid

What it misses

This policy does not state what it misses.

Examples from its eval suite

Warns: A README gains a Markdown image whose query sends a token to an outside host.

Allowed: A README gains shields.io and GitHub Actions badges without a query.

More

Built from engine 5a4b6287a46e and catalog 1e60d211c1bb.