Look up who has starred the microsoft/agent-academy repository on GitHub. Use this skill whenever the user wants to check stargazers, see if a specific GitHub user has starred the repo, or retrieve an up-to-date list of stargazers. Trigger whenever the user says "who starred agent academy", "check stargazers", "has <user> starred the repo", or asks anything about the repository's stargazer list. This skill uses the GitHub CLI to fetch stargazers via the GitHub API and saves the results to a local JSON file. It will install prerequisites if needed.

Generate a full Agent Academy feedback report — extracting feedback from Excel files and GitHub issues, analyzing sentiment, generating charts, and producing a single styled PDF with a cover page, management summary, and detailed analysis. Use this skill when the user asks to generate an Agent Academy report, create a feedback analysis, build a course completion report, or wants to analyze Agent Academy survey data. Also triggers when the user mentions Agent Academy feedback, course grades, sentiment analysis of Agent Academy data, or exporting Agent Academy results to PDF.

Update and maintain the llms.txt file. Use when the user wants to add sections, add links, correct information, or review the llms.txt file.

用于搜索金融市场、证券、上市公司基本面、价格、K 线、披露文件、财经新闻、A 股数据、港股或全球 ticker，可使用 yfinance、mootdx、API 脚本或 browser-use。

Hunt vector-store / embedding-layer weaknesses in RAG pipelines (OWASP LLM08 Vector and Embedding Weaknesses) — persistent corpus poisoning that survives across sessions and users (distinct from one-shot indirect prompt injection, which is owned by hunt-llm-ai), cross-tenant vector-database IDOR (unauthenticated or unscoped queries against Pinecone/Weaviate/Chroma/Milvus/Qdrant/pgvector), source-text/metadata leakage in similarity-search results, and retrieval-hijack via adversarial embedding proximity ('SEO poisoning' for RAG). Targets: any app with a shared knowledge base, document upload feeding a chatbot, or a directly reachable vector-DB port. Validate: a second, clean session/account must inherit a poisoned result, or a cross-tenant artifact must be independently verifiable — confabulation is not a finding, same bar as hunt-llm-ai. Use when target is RAG-backed, exposes a vector-DB port, or lets users upload documents that other users' queries later retrieve.

Hunt JWT cryptographic failures — alg:none signature-stripping and RS256→HS256 key-confusion that let an attacker forge a token for any identity (e.g. an admin) without knowing a secret. Use when the app authenticates with a JSON Web Token (an `eyJ...` Bearer token in the Authorization header, a cookie, or a login response). This skill OWNS JWT signature/crypto forgery (alg:none, key confusion, kid/jku header injection); hunt-ato covers JWT as one ATO path, hunt-auth-bypass covers SSO/SAML token trust, hunt-api-misconfig covers non-crypto JWT handling. Critical when a forged token grants access to another user's data or an admin-only endpoint.

Turn a completed daily UI inspiration capture into exactly five original landing-page builds, one per separate Codex task, using Sites. Use when the user asks to turn the daily inspiration references, a five-item UI prompt pack, or a dated `*-ui-inspiration-capture` article into distinct HTML landing pages while changing the source brands, names, copy, people, numbers, pricing, claims, and imagery.

Inspect and manage guarded Codex App-native or launchd heartbeats for Codex main control threads. Use when the user wants Codex to periodically continue a specific thread only when it is idle, avoid interrupting active work, keep App-visible interactions, inspect thread status, or enable/disable project autopilot heartbeats.

Prepare a new release for the C# MCP SDK. Assesses Semantic Versioning level (PATCH/MINOR/MAJOR), bumps the version, runs ApiCompat and ApiDiff, reviews documentation, updates changelogs, drafts release notes, and creates a pull request with all release artifacts. Use when asked to prepare a release, start a release, create a release PR, or assess what the next release should be.

Format code on the current branch using Biome. Use when asked to format, lint, or clean up code before committing or creating a PR.

Sandworm (APT44 / Seashell Blizzard, GRU Unit 74455) adversary-emulation playbook — IT→OT intrusion ending in ICS manipulation or destructive impact, executed with living-off-the-land Windows tooling. SAFETY-CRITICAL: destructive and ICS-write steps are canary/lab-only and gated on explicit OT authorization. Use when emulating Sandworm against an ICS/OT or critical-infrastructure estate. Triggers on: 'emulate Sandworm', 'APT44', 'Seashell Blizzard', 'Voodoo Bear', 'ICS attack', 'OT destructive', 'Industroyer', 'NotPetya'.

LockBit / generic RaaS-affiliate adversary-emulation playbook — broker/edge/RDP initial access, beacon, AD compromise to Domain Admin, defense evasion (Defender-disable via GPO, shadow-copy deletion), bulk exfil, then canary double-extortion encryption (Windows + ESXi). Reusable template for any ransomware affiliate (ALPHV, Akira, Black Basta). Triggers on: 'emulate LockBit', 'ransomware affiliate', 'RaaS', 'double extortion', 'StealBit', 'domain-wide ransomware', 'ESXi locker'.

Guides performance optimization for MSBuild engine code. Consult when working on hot paths in evaluation or execution, reducing allocations, choosing collection types, handling strings efficiently, modifying Expander.cs or Evaluator.cs, using Span<T>/stackalloc, caching values, or profiling build performance. Also applies when reviewing PRs for performance regression.

Guides changes to MSBuild binary log infrastructure. Consult when modifying BinaryLogger or BinaryLogReplayEventSource, adding new BuildEventArgs types, changing event serialization/deserialization, modifying ProjectImportsCollector, adjusting message importance levels, or making changes that affect .binlog content. Also applies when verifying that behavioral changes are properly reflected in binary log output.

Fix a Snyk vulnerability PR by regenerating the pnpm lockfile, checking changelogs for breaking changes, and posting findings as a PR comment. Use when asked to fix a vulnerability PR or handle a Snyk dependency upgrade.

Add a first-run guided tour, product walkthrough, coachmarks, or empty-state mock/example data to a Lightdash frontend feature. Use when the user wants to onboard users to a page, add a "Take the tour" flow, explain an unfamiliar UI, or show sample data on an empty page.

APT29 (Cozy Bear / Midnight Blizzard, SVR) adversary-emulation playbook — malware-light cloud-identity espionage: no-MFA password spray, OAuth consent/token abuse, Golden SAML, mailbox collection over residential proxies. Use when emulating APT29 against an M365/Entra/AWS-identity estate. Triggers on: 'emulate APT29', 'Cozy Bear', 'Midnight Blizzard', 'NOBELIUM', 'OAuth abuse', 'cloud identity espionage', 'Golden SAML'.

Review Azure SDK management-plane migration PRs (Swagger/AutoRest -> TypeSpec). Checks customization quality, TypeSpec decorator usage, and migration-specific anti-patterns on top of the standard mgmt PR review.

Backport commits from an existing GitHub Pull Request to a target branch and open a new backport PR. Use when the user asks to backport a PR, cherry-pick a merged PR into a release/maintenance branch, port a fix to an older version, or mentions "backport".

Audit the local dotnet/machinelearning ci-scan skill using recent ci-scan issues, maintainer feedback, and an optional local scan report, then draft targeted prompt improvements. Use when reviewing scanner quality, investigating false-positive KBEs, or improving the local ci-scan methodology.