Enforce security policies on AI agent tool calls and audit activity.
npx skills add https://github.com/PrismorSec/prismor --skill prismor
Please help me install this Agent Skill. Skill: prismor Source: https://github.com/PrismorSec/prismor/tree/main Command: npx skills add https://github.com/PrismorSec/prismor --skill prismor
Prismor addresses the security risks posed by AI coding agents by blocking dangerous actions, intercepting package installs, preventing secret leaks, and auditing session activity in real-time.
Install Prismor hooks into your project to enforce policy on agent tool calls:
prismor install-hooks --agent claude --workspace . --mode enforce
High-intent search queries and answers about installing and using this skill.
Runtime security for AI coding agents intercepts dangerous tool calls, blocks destructive commands, and substitutes real secret values with placeholders to prevent sensitive information from reaching the model.
Yes, you can audit package installations by evaluating them against threat intelligence before execution. The system records every tool call and maintains a tamper-evident audit trail of all agent actions.
You can enforce security policies by installing runtime hooks into your project. This intercepts prompt injection attempts and enforces tool call rules before the agent executes any code.
Runtime protection integrates with major AI coding agents and frameworks. It requires Python and provides a dashboard for monitoring activity, and generates signed evidence bundles for compliance auditors.
Shadow AI discovery secures environments by scanning the host machine to flag any AI agents running without security hooks. This ensures all autonomous coding sessions are monitored and policy-enforced.
The best way to block prompt injection attacks is deploying runtime security hooks that intercept and evaluate agent tool calls in real-time, blocking malicious prompts before they execute.
davidbeglenboyleExtract text and tables from PDFs into CSV or JSON using Python libraries.
haxlysEnforce Cedar policies and generate Ed25519-signed receipts for Claude Code tool calls.
sungaoxiang-backendExtract household registry image data into structured card_features slots.
AesopScottConfigure allowed-tool sets and execution policies for AI agent safety boundaries.