stbenjam

Compact conversation-steering skills for changing direction, explaining decisions, asking for clarity, answering questions without implied action, finishing punted work, delegating research to a subagent, tightening prose, executing decisively, rethinking code architecture, and raising frontend quality.

12 skills

Search and analyze a Calibre library or Goodreads export

11 skills

A grab-bag of miscellaneous OpenClaw skills (pollen forecasts, Orangetheory lookups, narrated video reels) with no unifying theme.

3 skills

Code review with specialist reviewers and runtime reproducers, plus guided eval and benchmark task walkthroughs

2 skills

Context-specific rules for safe Git operations and rigorous test-failure investigation.

2 skills

Skills for setting things up the way stbenjam likes them, such as giving every local agent the same skills and working principles.

2 skills

Hand-laid-out black-and-white box diagrams for design docs, rendered to SVG and PNG with a tiny Python shape DSL.

1 skill

Eliminate everyday friction through direct action, delegation, specialist services, or automation within explicit authorization boundaries.

1 skill

Injects a random motivational hype message into every turn via a UserPromptSubmit hook, plus about 10% of PreToolUse calls.

hooks

A Stop hook for Claude Code and Codex that requests one final completion pass, with an opt-in infinite mode.

hooks

Autonomous loops that shepherd work to completion, such as driving a PR to a mergeable state

1 skill