Agent Skills by Lysander Su
Showing 57 vetted skills indexed across 1 GitHub repositories.
fresh-context-review
Generates signed finding lists from PR diffs before formal code review.
organize-threads
Suggests label assignments for uncategorized threads by analyzing titles and metadata.
enterprise-workflow
Creates documents, tables, tasks, meetings, and slides in WeCom and Lark via callback APIs.
video-forge
Orchestrates end-to-end video production from asset ingestion through TTS dubbing, alignment, rendering, and review.
knowledge-engineering
Guides restructuring of external project documentation into structured, AI-indexable knowledge bases.
receive-handoff-grounding
Verifies handoff claims against independent resolvers before executing irreversible actions.
vision-rescue
Applies a five-step framework to break through seemingly impossible tasks before surrendering.
tech-writing
Writes narrative-driven technical articles from internal engineering practice for external audiences.
expert-panel
Orchestrates multi-agent expert debates with evidence chains and synthesized deliverables.
concept-demo-design
Designs interactive concept demos with typed contracts, delivery lanes, and verifiable evidence.
ttfund-skills
Query Tiantian Fund data through the official ttfund skills HTTP gateway.
thread-orchestration
Decomposes large tasks into parallel sub-threads with cat assignment and result aggregation.
co-creation-docs
Routes docs-only changes to direct push, pull request, or regular development based on risk classification.
context-self-management
Decides between handoff, sprint, or compression when context window warnings trigger.
eval-design
Designs and audits evaluation metrics using qualification gates, birth certificates, and health checks.
opensource-ops
Grounds, triages, and routes external PR and issue reviews with verified identity and custody rules.
code-as-harness
Diagnoses confirmed repeated user friction and proposes code-level harness fixes.
memory-navigation
Routes memory retrieval across search_evidence, graph_resolve, and list_recent entry points.
owner-friendly-plugin-development
Designs Clowder AI plugins with honest lifecycle states, in-product authorization, and verified release closures.
guide-authoring
Creates interactive onboarding guide flows as YAML with element tagging and registry discovery.
anime-forge
Produces AI-generated animated short films through keyframe, image-to-video, and editing pipeline stages.
proactive-memory-judgment
Evaluates person-related conversation cues to propose, defer, or abstain from memory proposals.
convention-graph-discovery
Maps repository conventions into a provenance-tracked graph of domains, extractors, and consumers.
memory-search-best-practices
Guides multi-query memory recall with per-question-type recipes and stopping criteria.