sickn33/agentic-awesome-skills v16.2.0

on GitHub

[16.2.0] - 2026-08-26 - "Security Operations Expansion and Safer Integrations"

Added a reviewed 43-skill security and reverse-engineering collection,
introduced API rate-limit, media-generation, and repository-instruction
workflows, corrected install guidance, and refreshed dependency safety. The
published catalog contains 2,074 skills.

This release expands defensive security, authorized assessment, reverse
engineering, incident-response, and threat-intelligence coverage while keeping
high-impact operations behind explicit scope, authorization, evidence, and
environment-safety gates.

Added

• Added 43 security skills adapted from the MIT-licensed
  zhaoxuya520/reverse-skill
  collection, spanning reverse engineering, malware analysis, forensics,
  threat hunting, API and identity security, firmware, wireless, cloud,
  supply-chain, and authorized assessment workflows
  (#1247).
• Added instructree for generating and validating
  scoped repository instruction trees without walking outside the selected
  project root
  (#1251).
• Added muapi-media for authenticated MuAPI image,
  video, and music generation with explicit cost, credential, and output
  handling
  (#1257).
• Added api-rate-limit-handler for
  retry-aware API clients with bounded backoff, jitter, idempotency guidance,
  and billable-request safeguards
  (#1258).

Changed

• Updated express-rate-limit from 8.6.1 to 8.6.2 in the generated Loki Mode
  example dependency set
  (#1256).
• Regenerated the canonical catalog, offline AAS Core data, tracked web assets,
  marketplaces, editorial bundles, compatibility reports, and Codex/Claude
  plugin distributions for 2,074 skills.

Fixed

• Corrected tree-ring-memory installation
  guidance to target the reviewed v0.15.0 release and aligned its declared
  risk with its install and update mutations
  (#1253).
• Sanitized embedded NUL bytes from imported Markdown in the security pack so
  the complete canonical skill subtrees remain text-safe and reviewable.
• Kept API retry guidance bounded and prevented automatic retries from silently
  multiplying non-idempotent or billable operations.

Security

• Excluded upstream runtime bootstrap scripts and the GPLv3 sidecar from the
  imported security collection; only reviewed Markdown skill content and
  references are distributed.
• Preserved explicit authorization, target-scope, lab-isolation, evidence, and
  stop-condition gates across offensive and dual-use security workflows.
• Classified credentialed media generation and repository-mutating install
  guidance as critical risk, with secrets supplied through environment or
  secure input rather than command-line arguments.