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 agentmods add skills/idaibin/skills/audit-javanpx skills add idaibin/skills --skill audit-javagit clone --depth 1 https://github.com/idaibin/skillsWhat 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 | $0.00067 | $0.01804 |
| Opus 5 | $0.00034 | $0.00902 |
| Sonnet 5 | $0.00013 | $0.00361 |
| Haiku 4.5 | $0.00007 | $0.00180 |
Grade A, and why
audit-java 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 2d 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 — 145 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Java Audit
Overview
Audit Java engineering from current repository evidence with explicitly selected risk profiles. Use open-source projects and framework guidance to form questions, not to declare target-repository defects by stylistic comparison.
Consume urn:skills:audit-request:v1; the portable output is
urn:skills:audit-findings:v1. When supplied, consume an
exact PackageManifest/basis plus compatible graph asset/consumer/impact results. The
audit produces findings and validation Observations, not Task/Requirement status or a
delivery Receipt.
Rule Priority
Apply current user intent, effective repository guidance, manifests/source/tests, project contracts, this Skill, then external references. A reference architecture never outranks a working local contract.
Workflow
- Record the Git/build root, revision and relevant Worktree state; read effective guidance and inspect only the selected manifests, source, tests, migrations, configuration ownership, CI, and runtime documentation.
- Resolve the JDK, Maven/Gradle owner, Wrapper, parent/BOM/platform, modules, packaging, framework generation, executable entries, profiles, and quality commands.
- Select one or more audit profiles:
- Build/architecture: dependency direction, modules/packages, public/internal seams, build reproducibility, dependency authority, generated code, and lifecycle.
- API/security: routes, DTO validation, filters/interceptors, authentication, authorization/data scope, sessions/tokens, CSRF/CORS, upload/download, errors, secrets, and sensitive logging.
- Persistence/transaction: mappings, query shape, pagination, N+1 risk, locking, migrations, transaction boundaries/propagation, after-commit effects, and database compatibility.
- Concurrency/integration: executors, async/events, Redis/cache, messages, schedules, retries, idempotency, backpressure, distributed locks, and shutdown.
- Performance/operations: representative workload, database plans, pools, allocation/serialization, caches, remote calls, metrics, health, and failure modes.
- Migration/compatibility: JDK, Spring Boot,
javax/jakarta, build tool, dependency, database, packaging, or configuration-generation transitions. Activate project grounding only for semantic signals involving runtime/config precedence, packaged artifacts, public contracts, durable data, legacy replacement, auth/security, or cross-repository delivery. Use the grounding chain to bound adjacent evidence; do not scan every profile or repository merely because corresponding files exist.
- Load Java engineering for every audit. Load codebase design for module/API/testability analysis and code quality only when maintainability is in scope.
- Consume a compatible graph asset/consumer/impact query or reproduce a bounded inventory of entries, callers, permissions, services, repositories, entities/mappers, migrations, configuration, jobs/listeners, and tests. Reject stale or mismatched graph results; a query miss never proves missing code, and a derived Markdown view is not audit evidence.
- Trace each candidate issue through trigger, reachable path, owner, state or data consequence, counterevidence, and a falsifiable validation seam. Reject checklist-only findings and style preferences without concrete impact.
- Capture tracked Worktree state before validation. Do not run known apply/fix or
tracked-source generation commands in the audited Worktree. If the repository offers
only write-capable validation, skip it and report the gap or hand it to a workflow
with an isolated Worktree. Prefer check-only repository commands and representative
data. With explicit authorization,
use only test-owned ephemeral containers, databases, brokers, or processes; never
write shared, staging, or production state. Recheck status/diff afterward. If an
otherwise non-mutating command creates tracked drift, stop validation, report the
exact contamination, mark affected evidence
Not verified, and do not revert it without explicit authorization. Clean up only task-created runtime resources. Compilation alone does not prove authorization, rollback, migration, query, concurrency, or runtime behavior. - Stop when the selected profiles are supported or explicitly blocked. Mark all other profiles out of scope.
- When Forgeway delivery integration is active, bind the audit to an immutable Run, exact input/result PackageManifest, and typed input refs. Attach each finding or validation result as an Observation against that package. Do not hand-edit a Gate, rewrite prior Observations after a retry, or infer reviewed/delivered state.
What ships with it
8 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
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.
- 2d ago First seen · 145 lines · 67 tokens per session scan A 501fddce2fa8
audit-java is a skill published in the GitHub repository idaibin/skills (6 stars, last pushed 5d ago), licensed Apache-2.0. It adds 67 tokens to every session and 1,804 once invoked, about $0.0003 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
airflow-java-sdk
Guide for contributing to the Airflow Java SDK (AIP-108). Use this skill whenever a contributor is working in the java-sdk/ directory or on the Java coordinator in task-sdk/src/airflow/sdk/coordinators/java/ — whether they want to add a feature, write tests, fix a bug, understand the architecture, or prepare a PR.…
jni-type-conversion
How to use @JniType annotations for ergonomic JNI. Relevant for Java files that use @NativeMethods or @CalledByNative.
wxjava-module-selector
根据微信公众号、小程序、微信支付、企业微信、开放平台、视频号或微信小店、腾讯企点和微信智能对话等业务场景,为用户选择合适的 WxJava Maven 模块、BOM 和示例入口。适用于用户询问“该用哪个模块”、依赖坐标、产品边界或单/多账号 Starter 选择时。.
azure-ai-formrecognizer-java
Azure AI Document Intelligence SDK for Java (com.azure:azure-ai-documentintelligence). Use for extracting text, tables, key-value pairs from documents, receipts, invoices, IDs, or building custom document models. Triggers: "document intelligence java", "form recognizer java", "extract text from PDF java", "OCR…
azure-ai-anomalydetector-java
Build anomaly detection applications with Azure AI Anomaly Detector SDK for Java. Use when implementing univariate/multivariate anomaly detection, time-series analysis, or AI-powered monitoring.
azure-communication-chat-java
Build real-time chat applications with Azure Communication Services Chat Java SDK. Use when implementing chat threads, messaging, participants, read receipts, typing notifications, or real-time chat features.