docs: update every HD path example/docstring to the 7-level hub-scoped shape (musehub#221)
Comprehensive sweep prompted by: has every place documenting or showing an example identity HD path actually been updated since #221 added the hub segment? Searched the whole ~/ecosystem tree for the MUSE_PURPOSE constant (1075233755) and for symbolic path-formula mentions (role'/index', domain'/entity_type', etc.) and checked every hit.
Deliberately left unchanged (correct as-is): - muse/core/hub_scoping_migration.py, muse/cli/commands/migrate_cmd.py, muse/cli/commands/path_cmd.py -- these legitimately show the six-level shape, since their whole job is detecting/migrating/decoding it. - muse/core/hdkeys.py's muse_path() docstring -- already correctly conditional (six-level when hub=None, seven-level when given).
Fixed (were stale six-level examples presented as current, or otherwise inconsistent): - docs/agent-provenance.md -- human + agent key examples now show the hub segment, with a pointer to docs/reference/auth.md for the full path breakdown. - docs/reference/auth.md -- both identity.toml examples now hub-scoped, plus a new explanatory paragraph on what hub' is and why six-level entries need muse migrate hub-scoping first. - muse/core/slip010.py -- this is the authoritative "Muse HD path structure" diagram (the module docstring every other doc effectively points back to). Rewrote it to show hub' as the optional, identity-domain-only segment, with an explanation of why other domains never use it. Updated the module's own runnable-looking examples to match. - muse/core/hdkeys.py -- one docstring assert used a non-computable "<hub>" placeholder where its sibling asserts show real literal values; replaced with the actual computed hub_index("musehub.ai"). - muse/cli/commands/auth.py -- _parse_rotation_index's docstring called the rotation index "the 6th path component" and showed a 6-level example; the function itself was already correct (takes parts[-1] regardless of length) but rotate now requires hub-scoped paths, so the doc claim was actively misleading about which position matters. - muse/cli/commands/agent.py -- three msign_path examples in the JSON schema docs were stale on two counts: they never accounted for hub scoping AND used a wrong abbreviated 4-field shape that never matched muse_path()'s real 6/7-field output even before #221. Replaced with the real computed shape matching the actual muse_path(DOMAIN_IDENTITY, ENTITY_AGENT, account, hub=hub_idx) call at the real call site.
Also fixed presentations/build-with-muse/episode-04 (separate repo, companion commit there) -- its identity.toml example was doubly stale: six-level AND showing the pre-migration fingerprints for gabriel's real identity, which changed when this same identity was actually migrated to hub-scoped keys earlier in this work. That script's own production notes explicitly require every shown artifact to be real and current, never a hand-typed placeholder -- pulled the actual current ~/.muse/identity.toml and a real verify-commit key_id to fix it properly rather than just patching the path shape cosmetically.
Tests: 206/206 across hdkeys/slip010/auth-rotate (docstring/comment-only changes, no functional code touched).
Semantic Changes
9 symbols
Files Changed
~6
1196 in snapshot
0 comments
muse hub commit comment sha256:b88ca1c72335d541266c2e7af62cc4235f3cd34191e62b5d3c5470beaf0a9d49 --body "your comment"
No comments yet. Be the first to start the discussion.