Explorez des publications vérifiées sur X, les personnes qui les partagent et les projets qu’elles mentionnent.
Seules les publications sur les Skills recensées et validées sont comptées. Une journée vide signifie qu’aucune publication n’a été recensée ici. Dates en UTC.
35 publications recensées
yara-l-authoring converts a threat description, ATT&CK technique, or Sigma, SPL, or KQL rule into a YARA-L detection rule plus positive and negative tests, // VERIFY field markers, and false-positive sources.
Marketing Skills hands coding agents a marketing playbook: landing-page conversion audits, copywriting, SEO and AI-answer visibility, analytics setup checks, email sequences and ad campaigns, all MIT-licensed.
php2shell-migration finds PHP apps exposed to PHP2Shell CVEs or stale runtimes, plans multi-framework upgrades on a separate Git branch, runs layered tests and reports exactly what was verified.
security-audit runs a six-phase review of a whole app inside Claude Code or Cursor, with a second agent checking every finding.
Universal Modder's mod-any-game skill has a coding agent identify a PC game's engine, read its real code, build the mod, test it in-game and record what it learned.
security-audit audits codebases and APIs with adversarial validation: a fresh agent tries to refute each finding, verdicts are confirmed, needs-evidence or rejected; target code only runs in an offline sandbox.
The offensive-active-directory skill packages Active Directory attack methodology for authorized red-team engagements, one of 22 attack surfaces in a pack that also covers SQL injection and phishing payloads.
am-i-cooked answers 'am i cooked?' for a coding agent: scan lockfiles in 18 formats, match exact versions against live CVE data, then fix on a branch and re-scan.
Safeguard, free for every AI agent, checks whether messages, links or pasted commands are scams and audits agent setups, hooks, secrets and dependencies as worms hit npm, PyPI and Composer.
security-audit runs a full codebase audit where one agent hunts vulnerabilities and a fresh agent tries to disprove each finding, so only survivors reach you.
rails-query gives coding agents a safe way to query live Rails data — counts, lookups, schema and EXPLAIN plans — locally or against production via Kamal, Heroku, SSH, Kubernetes or Docker.
phpcheck-url checks which PHP version a site advertises from its hostname, helping an agent spot installs that need patching.
check-side-effects sweeps a finished change for unintended security holes, breaking changes, slow paths and weakened checks; companion skills exercise UI, API and DB surfaces and prune tests that catch nothing.
security-audit reviews codebases, APIs and services for vulnerabilities before code ships, scaling from a focused review to a full audit or pen-test.
agent-blindfold keeps passwords and 2FA codes out of the model context: the user types them into a separate window, and a script fills the browser login.
production-dev-team makes Claude Code, Codex and Cursor agents ask before deleting data, pushing, deploying or running migrations, and treats nothing as verified without a recorded check.
user-behavior-qa drives a running app through its visible browser UI like a real user, then reports what works, what broke and what went untested without reading the code.
diffglow visual-diffs existing Playwright and Maestro screenshots, so an agent can catch UI changes that passing E2E tests miss, inspect them, then fix and rerun locally.
linkedin-bookmark-writer drafts, rewrites and critiques LinkedIn posts meant to be saved and shared, turning raw notes, a transcript or an article into founder, CIO or developer content.
VulnVerify bundles triage, poc-builder and fp-check Skills that verify web vulnerability reports from manual or agentic testing across OpenCode, Codex and Claude Code.