Loading...
Loading...
Found 14 Skills
Performs deep Root Cause Analysis (RCA) on NVIDIA TAO Visual ChangeNet classification experiments with image-evidence-driven investigation. Use when analyzing ChangeNet model failures, investigating poor recall / FAR / PASS-NO_PASS metrics, auditing visual inspection pipeline quality, or running an RCA report for an AOI defect-detection model. Trigger phrases include "RCA on my ChangeNet model", "why is my AOI model failing", "audit ChangeNet predictions", "investigate FAR regressions", "root cause analysis on visual-changenet".
Audit a whole regression suite and prune/restructure it with evidence: per-test coverage fingerprinting, AST near-duplicate clustering, CI-history mining for never-failing and flaky tests, prune decision rules (redundant/obsolete/low-value/keep), smoke/core/extended tiering by risk and defect-detection history, and a defensible "what we deleted and why" record. Deletion is destructive — quarantine and human sign-off are mandatory. Use when: "audit the test suite," "prune redundant tests," "find duplicate tests," "which tests can we delete," "restructure into smoke/core/extended," "is this test pulling its weight," "shrink the regression suite." Not for: Judging whether an individual test is WELL-WRITTEN (smells, assertions) — that is ai-qa-review. Healing one flaky test at runtime — that is test-reliability. Bulk selector regeneration after a UI refactor — that is selector-drift-recovery. Related: ai-qa-review, coverage-analysis, test-reliability, risk-based-testing, qa-project-context.
Optical Inspection for defect detection using Siamese networks. Compares image pairs to detect manufacturing defects, anomalies, or quality issues. Use when training, evaluating, exporting, or running inference for a TAO Optical Inspection model on AOI / quality-control data. Trigger phrases include "train optical inspection", "AOI defect detection", "Siamese defect classifier", "PCB / manufacturing inspection".
Visual ChangeNet for binary image classification and segmentation in AOI defect detection. Use when training, evaluating, exporting, or running inference for PCB defect detection or visual inspection, comparing image pairs for PASS/NO_PASS classification, or producing change-segmentation masks. Trigger phrases include "train Visual ChangeNet", "ChangeNet classify", "ChangeNet segment", "AOI defect detection", "PCB inspection model".
Review a Ralph plan and bd graph for defects via a subagent.
Provide structured code review guidance for catching defects and improving quality. This skill should be used when the user asks to 'review this code', 'check for issues', 'PR review', 'code quality check', or wants systematic code evaluation. Keywords: code review, PR, pull request, quality, defects, security, maintainability, performance.
Choose the smallest valuable automated test evidence and prove it can detect the defect it guards. TRIGGER when: behavior-changing work needs a test decision; adding, changing, deleting, or reviewing tests; a test may be vacuous. SKIP when: only running existing checks or completing live post-ship verification (use test-completed-plan).
Judge one bounded code candidate for defects, broad review, or read-only behavior parity for one stateful refactor or rewrite. Focus on exact identities, credible failure mechanisms, adversarial validation, maintainability evidence, and an evidence-backed result.
Review a change for evidenced defects and risks without editing it.
Review a code change, diff, commit, branch, or pull request for actionable defects, regressions, security risks, compatibility breaks, and missing verification. Use when code review is the primary requested outcome. Keep the review read-only unless the user explicitly asks to implement fixes; do not use as a substitute for implementation or a specialized visual audit.
Read-only defect review of a codebase - deep context and business-rule capture, perspective-based sweeps (business logic, error handling, security, correctness, concurrency, data integrity, performance, config, test quality), reproduction of each finding, severity calibration, and a decision-ready audit report with recommended fixes that are described but never applied. The codebase is left byte-identical. Use this whenever findings are wanted without changes, such as "review this codebase", "audit it but do not touch anything", "code review", "PR review", "second opinion on this code", "security assessment", "due diligence", "is this safe to ship", "what is wrong with this code", "review only", "report only", "read-only review", "check it without changing anything". Use it equally for business-logic defects where nothing crashes but the result is wrong - a wrong total, discount, tax or balance, a skippable step, a replayable action, a quota that does not hold, a user seeing another user's data. Prefer this over anti-bug whenever the user wants a report rather than a repaired repo, when the code belongs to someone else, or when changes need approval first. If the user actually wants the bugs fixed, use anti-bug instead.
Review code for actionable defects. Correctness is the core; performance and security are optional sub-cases of the same engine. Anchors on agreements between participants across a boundary, forces a violating execution, and refutes every candidate before reporting. Use when asked to review changes, find bugs, audit a codebase, or check whether a fix is safe.