Review engineering deliverables for policy compliance and cross-domain inconsistencies.
npx skills add https://github.com/Qiuu2/algo --skill critic-qiuu2
Please help me install this Agent Skill. Skill: critic Source: https://github.com/Qiuu2/algo/tree/main/.claude/skills/critic Command: npx skills add https://github.com/Qiuu2/algo --skill critic-qiuu2
Engineering deliverables for DSP, acoustic, hardware, structural, and test domains often contain unsubstantiated claims, policy compliance violations, or cross-domain inconsistencies that slip through standard review processes, leading to irreversible mistakes, costly rework, or failed stakeholder approvals.
Use the critic skill to perform a full adversarial review of the attached DSP filter design document and flag any compliance violations or unsubstantiated claims before it is submitted to the engineering lead.
High-intent search queries and answers about installing and using this skill.
To enforce engineering compliance, perform adversarial review of deliverables using provenance gates C1-C10 and L-grade completeness requirements. This identifies critical flaws and policy violations before stakeholder approval, ensuring all claims are substantiated.
Adversarial review is a policy-compliant scanning process for engineering deliverables that detects unsubstantiated claims and cross-domain inconsistencies. You need it when finalizing DSP algorithms, acoustic designs, or hardware schematics to prevent costly rework and irreversible mistakes.
Verify cross-domain consistency by applying interface checks that validate alignment between interconnected subsystems, such as matching acoustic cavity volumes with structural internal dimensions, and ensuring DSP MIPS requirements fit processor capabilities to prevent integration failures.
Yes, you can detect fixed-point overflow risks and missing thermal design validation through domain-specific error scanning. This process automatically identifies common error patterns across acoustics, DSP, hardware, structural, and test domains during the review.
Yes, the review process enforces irreversible action restrictions to prevent unverified data from being locked into decisions. All verdicts include auditable reviewer attribution, providing full traceability for compliance tracking per POLICY-PROV-001 v1.7.
The best way to review a hardware prototype bringup guide is to perform a full adversarial review to catch unsafe JTAG connection sequences, missing ERC/DRC waivers, and unverified component derating before submitting it to the engineering lead or CTO.
toneron2Prepare fractal horn geometries for L-PBF metal additive manufacturing.
ek33450505Require explicit human approval before executing destructive operations.
csml-rpiConduct mesh sensitivity studies and regression analysis for CFD simulations.
Qiuu2Orchestrate specialized AI agents through governed DAG workflows for audio hardware R&D.
