AGENT SKILL

security-check

Assess security implications of a software change using project scanners, dependency and secret checks, and changed trust boundaries. Use for requested security checks or changes affecting authentication, permissions, untrusted input, or sensitive data.

License 许可
MIT
最近更新
2026-11-10
Stars
★ 0
Forks
0

01 前言 Front Matter

name
security-check
description
Assess security implications of a software change using project scanners, dependency and secret checks, and changed trust boundaries. Use for requested security checks or changes affecting authentication, permissions, untrusted input, or sensitive data.

02 资源文件 1

03 SKILL.md 正文

Security check

Resolve the project, change range, supported languages, deployment context, and existing security configuration. Inspect changed trust boundaries: authentication and authorization, untrusted input, process invocation, filesystem access, data exposure, dependency changes, and workflow permissions. Use the repository's existing tools first. Do not turn a routine change into an unrelated full audit.

Choose evidence appropriate to the change:

  • Static analysis: use configured CodeQL or the project's equivalent. A completed scan is not a clean verdict; inspect findings and their affected commit. Model known exploit preconditions.
  • Dependencies: inspect lockfile changes and the package manager's audit or dependency-review results. A scanner with no supported dependency inventory has a coverage gap, not proof of safety.
  • Secrets: use configured secret scanning with redacted output. Keep values out of reports; record location and finding ID. If a real credential was exposed, report the need for rotation and controlled history remediation rather than merely deleting the current line.
  • Runtime behavior: use focused negative tests for changed permissions, validation, and failure paths. Keep probes within authorized targets and use owned fixtures.

Use project-check to capture local scanner commands under category security when available. For hosted scans, record the exact head SHA, scanner/job URL, completion state, finding dispositions, and unsupported or inaccessible checks. Use normal PR events with minimal workflow permissions; untrusted PR code must not execute with privileged secrets through pull_request_target.

Classify actionable findings by impact and reproducible evidence. Fix within scope, add an effective regression test where appropriate, then rerun the affected check. Explain false positives with specific evidence; retain existing thresholds instead of disabling rules to obtain a green check. Report passed, failed, blocked, and untested coverage separately. Scanner success alone does not establish that the application is secure.

When adding CI, use the CI reference. Install or configure account-bound tools only within the user's request; preserve existing authorization and repository policy.

04 同仓库更多 Skills

BrianLiDesign
BrianLiDesign/blidesign-ai-toolkit

toolkit-smoke-test

Verify a toolkit update across installed agent hosts using skill discovery, installed versions, MCP startup, and harmless tool calls. Use after installation or for cross-host release checks.

★ 0
BrianLiDesign
BrianLiDesign/blidesign-ai-toolkit

research-and-recommend

Research an open question or unfamiliar implementation, compare viable approaches against the user's constraints, and recommend the best fit using primary sources and applicable standards. Use when the user wants evidence-backed options and a recommendation; a single documentation lookup belongs to find-docs.

★ 0
BrianLiDesign
BrianLiDesign/blidesign-ai-toolkit

release-acceptance

Verify a deployed release against public behavior and persistence requirements, recording inspectable evidence and cleaning up owned test data.

★ 0
BrianLiDesign
BrianLiDesign/blidesign-ai-toolkit

project-check

Discover and run a project's verification commands, recording commit-specific results and missing checks. Use to validate a change or establish the repository's verification gate.

★ 0
BrianLiDesign
BrianLiDesign/blidesign-ai-toolkit

pr-closeout

Finish an authorized GitHub PR through review remediation, CI, metadata, merge, and post-merge verification. Use for PR closeout; keep assessment-only requests read-only.

★ 0
BrianLiDesign
BrianLiDesign/blidesign-ai-toolkit

orchestrate

Coordinate native subagents for substantial independent research, review, or implementation tasks, with bounded delegation, isolated edits, and verified integration. Keep small or tightly dependent tasks in the main agent.

★ 0