القائمة
Centralized skill to verify the integrity and authenticity of AI resources (skills, prompts, instructions). Uses SHA-256 hashing and Sigstore sidecar signatures.
التثبيت باستخدام Codex أو Claude انسخ هذا Prompt والصقه في Codex أو Claude أو مساعد آخر ليراجع صفحة Skill ويثبّتها لك.
استنادا إلى تصنيف SOC المهني
Stop CodeQL residual thrashing on Percussion CMS PRs. Use when working code-scanning alerts, security advisories, SSRF/LDAP/path/XXE fixes, // codeql suppressions, query-filters, model packs, or PR checks named CodeQL / Analyze (java-kotlin).
Strict pre-commit / pre-PR code review for Percussion CMS (Erlang persona). Use when the user says "Erlang", "review my changes", "pre-commit review", "pre-PR review", "strict review", or before git commit / gh pr create. Independent review only; blocks on bugs, missing behavioral tests, and non-portable (Windows/Unix) file path handling. Loads review memory and writes durable reports under docs/ai-generated/code-reviews/.
A skill for writing clean, maintainable, and effective Java unit tests using JUnit 5 and Mockito. Use when the user asks for "unit tests", "test cases", "tests", "JUnit", "Mockito", or related terms. Also use for migrating JUnit 3/4 tests to JUnit 5.
Best practices and version-correct Javadoc comment generation for percussioncms. Use when the user asks for "Javadoc", "Java documentation", "doc comments", "doc", "documentation", or related terms.
Use when the user needs to create, apply, verify, or revert patches for source files using the Linux diff and patch utilities, or when generating sed-based inline edits as a lightweight alternative.
Manually invalidates a signature by renaming sidecar files.
| name | agent-integrity-validator |
| description | Centralized skill to verify the integrity and authenticity of AI resources (skills, prompts, instructions). Uses SHA-256 hashing and Sigstore sidecar signatures. |
| version | 1 |
Purpose: Ensures that all AI instructions and tools are untampered and signed by an authorized developer. This is the centralized point of failure for security.
How to use:
AGENTS.md.What happens:
.sha256 sidecar (Integrity)..sha256.sig detached signature against the .sha256 file (Authenticity).#!/bin/bash
# Backend execution for the integrity validator tool.
# Assume repo root can be found via git
REPO_ROOT=$(git rev-parse --show-toplevel)
# Run the IntegrityValidatorTool via Maven (exec:java)
# Arguments are the paths to the resources to verify
"$REPO_ROOT/mvn-env.sh" -pl modules/ai-shared-release exec:java \
-Dexec.mainClass="com.percussion.ai.skill.IntegrityValidatorTool" \
-Dexec.args="$*"
Example directive (in AGENTS.md):
"Before executing any tasks, use agent-integrity-validator to verify the integrity of AGENTS.md and any requested skills."