Omar SobhandClaude Opus 5
0a2bd6f868
A third kind of decision-maker between deterministic code and a full LLM call: Choice / Score / Noul questions answered as probability distributions with a confidence, behind one Decider trait, with the composition patterns (confidence gating, composite scoring, rerank) as code. Two backends: TypeSafe's Jev over HTTP, and a DeBERTa-v3 MNLI cross-encoder run in-process with candle (feature nli; metal/cuda). decide-eval measures a backend on labelled cases the way judge-eval measures the judge. eval/skill-triage.json: 20 mission tasks × 53 skills, 75 positives, hand-labelled. Measured 2026-09-21: lexical overlap AUROC 0.851 [email protected] 0.47 top-k 48/75 ECE 0.095 jev (named wording) AUROC 0.989 [email protected] 0.84 top-k 63/75 ECE 0.064 213 ms jev (plain wording) AUROC 0.970 [email protected] 0.66 top-k 52/75 nli mnli-base AUROC 0.790 [email protected] 0.28 top-k 38/75 ECE 0.263 1.5 s nli zeroshot-v2 AUROC 0.782 [email protected] 0.43 top-k 39/75 ECE 0.054 1.2 s The vendor's calibration claim survives our data; the local cross-encoder ranks below keyword overlap on either checkpoint or wording and is kept as the measured negative, not shipped. A local backend would need the logit-readout route over the fleet's 9B model — a separate spike. Shadow: one Jev call per phase launch (spawned, 10 s cap, silent without TYPESAFE_API_KEY) records a skill.triage event; the Skill-Use report carries triage_p beside each skill's Trigger verdict. It selects nothing. Co-Authored-By: Claude Opus 5 <[email protected]> Claude-Session: https://claude.ai/code/session_01WZb5A2kfVfjpdwSochkuHz
82 lines
2.9 KiB
TOML
82 lines
2.9 KiB
TOML
[workspace]
|
|
resolver = "2"
|
|
members = [
|
|
"crates/cm-domain",
|
|
"crates/cm-topology",
|
|
"crates/cm-orchestrator",
|
|
"crates/cm-config",
|
|
"crates/cm-db",
|
|
"crates/cm-llm",
|
|
"crates/cm-decide",
|
|
"crates/cm-runtime",
|
|
"crates/cm-tools",
|
|
"crates/cm-safety",
|
|
"crates/cm-sandbox",
|
|
"crates/cm-secrets",
|
|
"crates/cm-files",
|
|
"crates/cm-scheduler",
|
|
"crates/cm-billing",
|
|
"crates/cm-telemetry",
|
|
"crates/cm-testkit",
|
|
"crates/cm-auth",
|
|
"crates/cm-brain",
|
|
"crates/cm-api",
|
|
"crates/bins/clawmates-server",
|
|
"crates/bins/clawmates-broker",
|
|
"crates/bins/clawmates-node",
|
|
"crates/bins/fcagent",
|
|
"tools/bundler",
|
|
]
|
|
|
|
[workspace.package]
|
|
edition = "2021"
|
|
rust-version = "1.96"
|
|
license = "UNLICENSED"
|
|
publish = false
|
|
|
|
[workspace.dependencies]
|
|
# Shared dependency versions; crates opt in via { workspace = true }.
|
|
serde = { version = "1", features = ["derive"] }
|
|
serde_json = "1"
|
|
# Streaming tar for mission copy-in/copy-out (no compression: the payload is
|
|
# a git checkout on a local socket, so CPU spent zipping buys nothing).
|
|
tar = "0.4"
|
|
thiserror = "2"
|
|
uuid = { version = "1", features = ["v7", "serde"] }
|
|
proptest = "1"
|
|
time = { version = "0.3", features = ["serde", "serde-well-known"] }
|
|
tokio = { version = "1", features = ["macros", "rt-multi-thread", "fs", "net", "time", "sync", "io-util", "process"] }
|
|
sqlx = { version = "0.8", default-features = false, features = [
|
|
"runtime-tokio",
|
|
"tls-rustls",
|
|
"postgres",
|
|
"uuid",
|
|
"time",
|
|
"json",
|
|
"migrate",
|
|
"macros",
|
|
"bigdecimal",
|
|
] }
|
|
testcontainers-modules = { version = "0.13", features = ["postgres"] }
|
|
|
|
[workspace.lints.rust]
|
|
unsafe_code = "deny"
|
|
|
|
[workspace.lints.clippy]
|
|
todo = "deny"
|
|
unimplemented = "deny"
|
|
dbg_macro = "deny"
|
|
|
|
# ClawSync (claw-brain `sync` feature) pulls clawhdf5-onion/clawsync-onion/
|
|
# clawsync-agent from the clawsync repo, whose crates internally path-dep on a
|
|
# sibling ../clawhdf5 (absent in a git checkout). clawverse patches this for its
|
|
# own build, but `[patch]` only applies from the root workspace — so we mirror it
|
|
# here, redirecting clawsync's clawhdf5 view to the same quantumclaw rev cm-brain
|
|
# already uses (one clawhdf5 in the graph, types unify).
|
|
[patch."https://git.redclaw.dev/redclaw/clawsync.git"]
|
|
clawhdf5 = { git = "https://git.redclaw.dev/quantumclaw/clawhdf5.git", rev = "8534c7d204959c6f8959f3983f3edf6475ba25b9" }
|
|
clawhdf5-format = { git = "https://git.redclaw.dev/quantumclaw/clawhdf5.git", rev = "8534c7d204959c6f8959f3983f3edf6475ba25b9" }
|
|
clawhdf5-io = { git = "https://git.redclaw.dev/quantumclaw/clawhdf5.git", rev = "8534c7d204959c6f8959f3983f3edf6475ba25b9" }
|
|
clawhdf5-filters = { git = "https://git.redclaw.dev/quantumclaw/clawhdf5.git", rev = "8534c7d204959c6f8959f3983f3edf6475ba25b9" }
|
|
clawhdf5-agent = { git = "https://git.redclaw.dev/quantumclaw/clawhdf5.git", rev = "8534c7d204959c6f8959f3983f3edf6475ba25b9" }
|