Results for “weakly-supervised”

Flamingo: A Visual Language Model for Few-Shot Learning

Enables decoupled pub/sub messaging between ViewModels using CommunityToolkit.Mvvm Messenger, covering weak/strong references, request/reply patterns, channel tokens, and lifecycle management.

Skeptical pushback on visual placement — both density and quality. Reads the annotated outline plus the visuals manifest and asks (a) whether the article hits the density target from editorial-principles-visuals.md, (b) whether each [VISUAL:...] earns its place, (c) whether sections without one would benefit. One revision pass on FAIL (BLOG_AGENT_VISUALS_REVISION_BUDGET, default 1).

CogVLM: Visual Expert for Pretrained Language Models

Open-Vocabulary Object Detection Using Captions

Expert guidance for distributed training with DeepSpeed - ZeRO optimization stages, pipeline parallelism, FP16/BF16/FP8, 1-bit Adam, sparse attention

WOLVERINE v2.0 — HYPE alpha hunter. Entry-only scanner, DSL-exit-only architecture. v1.1 lost -22.7% because the scanner's thesis exit chopped 25/27 trades before DSL could manage them. v2.0 removes thesis exit entirely. Scanner decides entries (score 8+, 4H/1H aligned, SM consensus). DSL manages all exits (wide Phase 1 for HYPE volatility, trailing tiers starting at +15% ROE). Leverage lowered to 7x. Max 4 entries/day. 3-hour cooldown between entries. DSL exit managed by plugin runtime via runtime.yaml.

PaLI-3: Smaller, Faster, Stronger

Audits project dependencies for supply chain risks including single maintainers, unmaintained packages, low popularity, high-risk features, past CVEs, and missing security contacts.

Identifies the weakest samples per ground-truth label in NVIDIA TAO VCN Classify experiments by running a Docker container that performs threshold sweep, weakness scoring, and per-lighting expansion, then surfaces top-K weak samples for downstream augmentation or relabeling.

Build, deploy, evaluate, debug, and measure latency for a GPU-accelerated FastAPI service that detects whether operators perform assembly-line steps in order via event boundary detection and VLM classification.

Operate the VSS alert pipeline for real-time monitoring, Alert-Bridge subscriptions, Slack notifications, incident queries, and camera onboarding.

Analyze the available skill set and recommend candidate groups that could be consolidated into sub-skill bundles. Read-only — proposes a plan, does not move files.

Multimodal Few-Shot Learning with Frozen Language Models

Security vulnerability scanner and OWASP compliance auditor for codebases. Dependency scanning (npm audit, pip-audit), secret detection (high-entropy strings, API keys), SAST for injection/XSS vulnerabilities, and security posture reports. Activate on 'security audit', 'vulnerability scan', 'OWASP', 'secret detection', 'dependency check', 'CVE', 'security review', 'penetration testing prep'. NOT for runtime WAF configuration (use infrastructure tools), network security/firewalls, or compliance certifications like SOC2/HIPAA (legal/organizational).

Audits OAuth 2.0 permissions across identity providers to identify over-privileged third-party integrations, excessive API scopes, and unused token grants, enforcing least-privilege access.

Improved Baselines with Visual Instruction Tuning

WOLF v6.3 — Fully autonomous multi-strategy trading for Hyperliquid perps via Senpi MCP. Manages multiple strategies simultaneously, each with independent wallets, budgets, slots, and DSL configs. 5+N cron architecture: 5 shared wolf crons (Emerging Movers 3min, SM Flip 5min, Watchdog 5min, Risk Guardian 5min, Health Check 10min) plus one DSL v5.2 cron per strategy (native Hyperliquid SL sync via dsl-dynamic-stop-loss skill v5.3.1). Same asset can be traded in different strategies simultaneously. Enter early on first jumps, not at confirmed peaks. Dynamic risk-based leverage per strategy. Requires Senpi MCP connection, python3, mcporter CLI, OpenClaw cron system, and dsl-dynamic-stop-loss skill (provides dsl-cli.py + dsl-v5.py).

Harden Kubernetes Role-Based Access Control by implementing least-privilege policies, auditing role bindings, eliminating cluster-admin sprawl, and integrating external identity providers.

Audit and harden AWS Lambda execution roles by implementing least-privilege IAM policies, permission boundaries, and SCP enforcement.

Simulates bandwidth throttling and network degradation attacks using tc, iperf3, and Scapy in authorized environments to test quality-of-service controls, application resilience, and network monitoring detection of traffic manipulation attacks.

Screen US equities for parabolic exhaustion patterns, generate conditional pre-market short plans, and monitor intraday trigger signals using 5-min bars.

Accelerate LLM inference using speculative decoding, Medusa multiple heads, and lookahead decoding techniques for 1.5-3.6× speedup without quality loss.

Visual Instruction Tuning

DEPRECATED in v5.6.0 — see joint-multi-tf-v560 skill. Documents the v5.2.0 dual-model approach (train separate 15Min/1Hour models, combine via weighted voting). Still relevant for: (1) loading legacy v5.5.0 dual models, (2) understanding the historical aggregation layer, (3) resampling pattern via origin='start'.

Track ML experiments with automatic logging, visualize training in real-time, optimize hyperparameters with sweeps, and manage model registry with W&B.

Scaling Vision with Sparse Mixture of Experts

Audit Kubernetes RBAC configurations to identify overly permissive roles, wildcard permissions, dangerous bindings, service account abuse, and privilege escalation paths using kubectl, rbac-tool, KubiScan, and Kubeaudit.

Audit agent skills, plugins, prompts, manifests, scripts, dependencies, and bundled assets for provenance, prompt-injection, permission, execution, exfiltration, persistence, and update risk. Use when evaluating a third-party skill before installing, enabling, updating, publishing, or distributing it; reviewing an untrusted SKILL.md, agent configuration, MCP integration, archive, or repository; comparing a package with a known-good version; or investigating unexpected tool, network, credential, or filesystem behavior.

VS-Enhanced Theoretical Framework Architect with Critique & Visualization Full VS 5-Phase process: Modal theory avoidance, Long-tail exploration, differentiated framework presentation Absorbed A3 (Devil's Advocate) critique and A6 (Conceptual Framework Visualizer) capabilities Use when: building theoretical foundations, designing conceptual models, deriving hypotheses, critiquing frameworks, visualizing models Triggers: theoretical framework, 이론적 프레임워크, conceptual model, 개념적 모형, hypothesis derivation, critique, devil's advocate, 반론, visualization, diagram

Discover and reorganize the skill inventory into hierarchical sub-skill bundles. Use when the user asks to review, group, or consolidate skills into a parent bundle.

Use when simulating paired-end short reads from a reference FASTA for testing, benchmarking, or pipeline validation

SNLI-VE: Visual Entailment Dataset

BALD EAGLE v3.0 — XYZ Alpha Hunter (Hardened). Focused on 6 high-liquidity XYZ assets: CL, BRENTOIL, GOLD, SILVER, SP500, XYZ100. Conviction-scaled leverage (5-10x based on score). Wider DSL for macro assets. Maker-only execution. Scanner calls create_position internally. v3.0: focused assets, conviction-scaled leverage, XYZ-tuned DSL, no thesis exit.

Render long lists (1k+ rows) efficiently in Salesforce LWC using lightning-datatable infinite loading or an IntersectionObserver sentinel, with keyset pagination and accessibility guidance.