Loading...
Loading...
Found 57 Skills
Inspection visuelle et revue de code pour la conformité aux Web Interface Guidelines. Se déclenche sur « review my UI », « check accessibility », « audit design », « review UX », « fix the layout », « find design problems ». Supporte l'analyse statique du code et l'inspection visuelle dans le navigateur avec correction automatique.
Static inspection of Triton operator code quality (Host side + Device side) for Ascend NPU. Used when users need to identify potential bugs, API misuses, and performance risks by reading code. Core capabilities: (1) Ascend API constraint compliance check (2) Mask integrity verification (3) Precision processing review (4) Code pattern recognition. Note: This Skill only focuses on static code analysis; compile-time and runtime issues are handled by other Skills.
Code-level quality analysis. Use when asked to review code for smells, security issues, implementation quality, or test coverage.
Static source-code vulnerability scan. Reads a target directory (and THREAT_MODEL.md if present), spawns parallel review subagents per focus area, and writes VULN-FINDINGS.json + .md for /triage to consume. Read-only — no building, running, or network. For execution-verified crashes, use vuln-pipeline instead. Use when asked to "scan for vulns", "review this code for security issues", "find bugs in <dir>", or as the step between /threat-model and /triage.
Perform a static, read-only code review of an Astro pull request or of a local branch, commit range, diff, patch, or working tree being prepared as a pull request. Use this skill only when the user asks to review PR changes, review their current branch before opening or updating a PR, or self-review a PR for correctness, security, tests, simplicity, runtime portability, error handling, comments, behavior documentation, and changeset coverage. Do not use it for issue review, bug-report investigation, reproduction, diagnosis, or triage. This skill reports findings only: it never edits code or runs project code, tests, builds, checks, or scripts
Use when reviewing scripts, client components, bundles, or runtime behavior related to Lint JavaScript code. Inspect both source code and the browser execution path so fixes target the real bottleneck or bug.
Triage supplied or imported security findings against a repository using its security policy and static code evidence. Accepts scanner reports, advisories, GitHub findings, and Jira or Linear tickets. Do not use for discovery, duplicate triage, runtime validation, or fixes.
Classify a code quality concern into the right enforcement tool and act on it. Activate when the user wants to enforce a pattern, catch a mistake, add a check, create a rule, prevent a practice, guard against regressions, set up linting, improve their feedback loop, or asks "how do I make sure X."
Critical Standard Bottom-Up Self-Check (check-standards, Convergence Gate) after Java backend code generation/modification. **This skill must be loaded after any Java code is written or modified—including scattered coding that does not follow the complete development process (e.g., "write an XX interface"/"write a Controller"/"help me write this code"): this skill must be loaded and verified before delivering code products to users**. It uses grep/ast-grep to actually scan and verify each standard—comments (class/field/method Javadoc/step comments, including private and extracted methods), logs (full class @Slf4j, each business method has logs, ERROR exceptions with stack traces), transaction rollbackFor, constructor injection, layered boundaries, SQL in XML, unified return body, pagination limit, password encryption, SQL comments/injection/WHERE, and all other verification items, with "File:Line Number" evidence attached to each item; **No level distinction—any item that is not implemented properly (regardless of importance) will be listed in the pending confirmation list to confirm with the user whether to complete it**. **Correct intermediate product naming and paths before verification** (Technical Solution 3.x.1 / Interface List 3.x.2 / Standard Verification Report 5.2.x / Acceptance Report 5.3.x, remove the task ID prefix T0xx from file names and move to the module version directory). Trigger scenarios: after writing code, after code generation, after modifying code, **before delivering code**, pre-submission check, code standard verification/review, comment and log verification, code review, standard self-check, bottom-up check, acceptance review, product file naming correction, file renaming and relocation, **perform standard scanning on specified directories/modules/packages (e.g., "run check-standards on xx directory"/"verify order module standards"/"only scan the code modified this time")**. It can be used with ai-dev-workflow 5.2 standard verification node / 5.3 acceptance, or triggered independently (output the report in the dialogue and prompt the user for confirmation when there are no docs/no process).