alex_metacraft

4 ms·

1. 1.▲
   by alex_metacraft 2mo ago

   So cool! I gave it a try and used it for web design and presentations. It satisfied both of my needs and I gave up on the Pencil MCP immediately.
2. 2.▲
   Show HN: Avibe – your AI agent lives on your machine, reachable from your phone(github.com)

   4 points by alex_metacraft 4mo ago|1 comments
3. 3.▲
   by alex_metacraft 8mo ago

   Good catch on the numbers. 29/33 vs 33/33 is the kind of gap that could easily be noise with that sample size. You'd need hundreds of runs to draw any meaningful conclusion about a 4-point difference, especially given how non
4. 4.▲
   by alex_metacraft 8mo ago

   Hey HN. I've been working on askill, a CLI package manager for agent skills (SKILL.md files used by Claude Code, Codex, Cursor, etc.). There are already several skill directories and installers out there (skills.sh, skillregistry.io, a
5. 5.▲
   Show HN: Askill – A package manager for AI agent skills with AI safety scoring(github.com)

   1 points by alex_metacraft 8mo ago|1 comments
6. 6.▲
   by alex_metacraft 8mo ago

   I think this experiment has a fundamental flaw in its comparison setup. What they're comparing is: (A) a skill with a short description in the frontmatter, which the agent may or may not decide to invoke, vs. (B) a massive compressed i
7. 7.▲
   by alex_metacraft 8mo ago

   This is a really interesting finding. It makes sense when you think about what the training data looks like — first person statements in a system prompt pattern-match to "internal monologue" or "chain of thought" example