AGENT SKILL

threat_model_projection

Project adversarial candidates into declared norm balls and valid input boxes for robustness evaluation.

License 许可
Apache-2.0
最近更新
2026-09-03
Stars
★ 351
Forks
27

01 前言 Front Matter

name
threat_model_projection
description
Project adversarial candidates into declared norm balls and valid input boxes for robustness evaluation.

02 资源文件 2

03 SKILL.md 正文

Threat Model Projection

Use this skill when implementing or auditing adversarial attacks that must stay within an Linf or L2 perturbation budget. Do not use it to choose epsilon values or to evaluate model accuracy by itself.

Inputs

  • Clean vector or batch of vectors.
  • Candidate adversarial vector or batch.
  • Norm name: Linf or L2.
  • Epsilon and input bounds.

Outputs

  • Projected adversarial vector.
  • Diagnostics containing perturbation norms and clipping status.

Workflow

  1. Subtract clean inputs from candidates to obtain perturbations.
  2. Project perturbations to the requested norm ball.
  3. Add projected perturbations to clean inputs.
  4. Clip to the valid input box.
  5. Recompute diagnostics and assert invariants before using the result.

Validation

Run python tests/test_projection.py or validate through the bundled skill-tree validator.

Limitations

The script is pure Python for small deterministic recovery tests. Large tensor workloads should use a vectorized framework while preserving the same contracts.

04 同仓库更多 Skills

VectorSpaceLab
VectorSpaceLab/AREX-Skill

passnet-triton-opt

Make a single PassNet Triton kernel fast AND numerically faithful: performance model (when a replacement can win at all), block/grid/warp tuning, autotune policy, launch overhead, and per-op numeric recipes to pass the dtype baseline tolerances. Use when a pass matches and is correct but speedup below expected, or when correctness fails by small numeric margins.

★ 351
VectorSpaceLab
VectorSpaceLab/AREX-Skill

passnet-solve

END-TO-END playbook for solving one PassNet sample: analyze the computation graphs, decide the optimization strategy (what to fuse, what to replace, what to leave alone), drive the iteration loop, and maximize the sample score. This is the ENTRY skill — invoke it first for any PassNet optimization task; it tells you when to use passnet-pattern-fusion, passnet-triton-opt and passnet-feedback.

★ 351
VectorSpaceLab
VectorSpaceLab/AREX-Skill

passnet-skill

PassNet GPU kernel optimization via compiler passes. Design and implement Triton-based optimization passes, create pass files under ./pass_dir/, self-evaluate with pass_evaluator, and iterate to maximize GPU speedup.

★ 351
VectorSpaceLab
VectorSpaceLab/AREX-Skill

passnet-pattern-fusion

Author PassNet pass files: write patterns that actually MATCH the FX graph, pick fusion regions (which ops to absorb into one Triton kernel), and structure multi-pass submissions with the mandatory shared-dispatch architecture. Use when creating or fixing pattern()/replacement_args()/replacement_func() files, when a pass "failed to match", or when deciding how to fuse multiple kernels/ops.

★ 351
VectorSpaceLab
VectorSpaceLab/AREX-Skill

passnet-orchestrate

Round/batch planning and decision support for PassNet graph optimization. Use for multi-sample triage, worker allocation, eval-budget policy, and post-eval keep/revert decisions. For solving one concrete sample end to end, use passnet-solve as the entry skill; consult this skill only when you need broader planning or a specific decision gate.

★ 351
VectorSpaceLab
VectorSpaceLab/AREX-Skill

passnet-feedback

Fast, deterministic feedback for PassNet work: pre-flight pattern/match verification WITHOUT burning a GPU evaluation, per-node bottleneck analysis of a sample's graphs, and eval-log parsing into per-variant status + estimated score + failure classification. Use BEFORE every GPU evaluation (check_pattern), at the START of a sample (analyze_graph), and AFTER every evaluation (parse_eval_log).

★ 351