Source Intelligence

DisclaimerUnofficial, and not affiliated with Anthropic. Nearly all of this is read straight out of what ships: npm bundles, captured prompts, published docs. Anthropic's own notes go in verbatim, marked as theirs. The rest is my reading, and every entry carries the strings behind it. If one looks wrong, vote it down and say why.

All of v2.1.232 Home All releases olderv2.1.231 v2.1.233newer
Claude Code v2.1.232

Skill files can be read through a mounted memory store

Not switched on
Useful2 Signal3
Skills

Skills can be discovered from a mounted memory store, with reasons given when one is rejected.

Store-backed SKILL.md reading only applies when a memory store and project key are mounted; otherwise disk is used.

What

When a memory store and project key are available, skill discovery reads SKILL.md through that store instead of resolving it on disk, and reports why a skill was rejected rather than skipping it silently. Without a store mounted, the old disk path is used unchanged.

Details
  • Typed failures cover a file not at its apparent location, a file moved while being opened, a second name pointing at the same file (skipped for this cycle), and anything that is not a plain small file.
  • Symlink detection moved from calling isSymbolicLink() to reading a symlink field.
Evidence

is not a plain small file

Strings lifted out of the shipped bundle, so the claim above can be checked against them.

See this entry in the whole of v2.1.232 →