docs: document the +cric cryptosuite that replaced +crub in Ames and Jael

[sigilante]

Sixth PR from the audit against urbit/urbit@08026c84b2. Cross-cutting — this one was flagged independently by two passes of the audit, in two different doc trees, which is what pushed it up the queue.

Companion PRs: #266, #267, #268, #269, #270.

What changed upstream

+cric:crypto (zuse.hoon:1489) is the suite Ames and Jael now use. Unlike +crub, which is fixed to Suite B, +cric is a door over a tagged payload and can carry either suite:

$%  [suite=%b pub=[cry=@ sgn=@ ~] sek=...]
    [suite=%c pub=[cry=@ sgn=@ tw=[ugn=@ dat=@ xtr=@]] sek=...]
==

Counted rather than assumed: sys/vane/ames.hoon contains zero references to crub and 26 to cric. Jael builds keys through +cric too (pit:nu:cric:crypto, nol:nu:cric:crypto — jael.hoon:266,338,854,1260,1271,1306).

+crub still exists in zuse (:1703) and its documentation is kept — it's now labelled superseded rather than deleted, since the docs previously called it "the only cryptosuite which Ames makes use of".

One structural change worth calling out: the symmetric arms moved into a +cyf sub-core. What the docs called +en:crub:crypto is now +en:cyf:cric:crypto. All four sub-cores are documented: as, ex, nu, cyf.

Corrections that required reading the current state, not the migration code

Two things here are worth a reviewer's attention, because the obvious search gives the wrong answer.

1. Ames no longer stores a crypto core at all. The audit described Ames as wrapping _cric:crypto in an internal +acru-25 core with a legacy +acru-12 for old peers. That is true of the migration types only. The current +$ axle (lull.hoon:1601, axle-31) has no crypto-core field — it stores [saf=keypairs =ring =pass], where +$ keypairs is [pub=public-keys sek=private-keys] (lull.hoon:1069). +acru-12 (ames.hoon:817) and +acru-25 (:881) appear only inside versioned definitions like ames-state-13 and ames-state-22.

The docs now state this explicitly, so the next reader doesn't repeat the mistake. (This is the same shape of trap as the heeds field in #269.)

2. Comet self-attestation no longer signs via an acru core. +etch-open-packet (ames.hoon:423-438) signs with (sign-raw:ed:crypto - [sgn.pub sgn.sek]:saf). The documented +sign:as:crub was wrong in both suite and arm.

Not changed, having been checked

• jael/data-types.md $point and $key-update already carry crypto-suite=@ud and match lull.hoon:4323,4327. The audit listed these as needing the field added — they did not. Only $point's prose omitted an explanation of what crypto-suite means, which is now supplied.
• +crub:so (stdlib/4l.md) is an unrelated date-parsing arm in +so. Untouched.
• core-academy/ca02.md quotes +crub source, but Core Academy declares itself a snapshot at Zuse 412k.

All anchor links in the touched files were verified to resolve, and this branch was test-merged against the five open companion PRs — all clean.

🤖 Generated with Claude Code

[thelifeandtimes]

Deferring to @hanfel-dovned for this

[hanfel-dovned]

Looks good to me. The edge case I describe here may or may not be worth drawing attention to.

[hanfel-dovned]

Confidential comets do introduce an edge case where you might not have the entire history of a point's networking keys stored in its Jael state, but this should probably be discouraged at the very least, and depending on how we implement the namespace it may have to be forbidden.