CyberSecurityUPandClaude Opus 5 d334946915 bench: refresh the TypeSafe benchmark — 13/13 coverage, data-type-aware severity ...

Current-build run against the 13-scenario target with TypeSafe on: every seeded
class confirmed with a live receipt, chained beyond the set into full admin
takeover, GraphQL authz bypass, a config secret leak and an authenticated RCE.
The credential-dump BOLA holds Critical because severity is graded on the kind
of data exposed, not the class. report.html + run artifacts + README refreshed;
no secrets committed.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>

2026-09-20 14:06:39 -03:00

..

agents.rs

v3.6.0 — AI/LLM/Agent/MCP/Skills security, n8n audit, onboarding wizard

2026-07-10 11:09:19 -03:00

assurance.rs

feat(assurance): target gate (default-deny), evidence-graded CVSS, audit anchoring, P1–P5 bundle

2026-09-18 21:25:21 -03:00

attack_graph.rs

feat(chain,skills): close benchmark misses — CRLF-on-Location, second-order precondition; condense BENCHMARK

2026-09-20 12:42:36 -03:00

audit.rs

feat(assurance): target gate (default-deny), evidence-graded CVSS, audit anchoring, P1–P5 bundle

2026-09-18 21:25:21 -03:00

belief.rs

v3.6.5: LLM red-teaming (jailbreaks & prompt injection) + Opus 5 / Sonnet 5 / Kimi K3

2026-07-28 13:38:15 -03:00

browser.rs

feat: browser validator for XSS, and replay that separates request from effect

2026-09-14 00:32:31 -03:00

budget.rs

feat(budget,provenance): reasoning budget modes and JOASNSCOPE provenance

2026-09-14 01:14:03 -03:00

capability.rs

feat(harness): risk model, engagement policies, capability tokens, audit trail

2026-09-13 17:56:31 -03:00

chain.rs

feat(chain,skills): close benchmark misses — CRLF-on-Location, second-order precondition; condense BENCHMARK

2026-09-20 12:42:36 -03:00

claims.rs

feat(claims): separate mechanic from impact so an overstatement stops deleting the observation

2026-09-14 00:27:29 -03:00

compliance.rs

feat: PoC validator, Kali sandbox, intercept proxy, compliance, +8 validators

2026-09-14 10:26:06 -03:00

creds.rs

identification/attribution + multi-role access-control auth (v3.5.5)

2026-07-01 23:59:02 -03:00

cvss.rs

feat(assurance): target gate (default-deny), evidence-graded CVSS, audit anchoring, P1–P5 bundle

2026-09-18 21:25:21 -03:00

grounding.rs

feat: human-in-loop validator (flag not delete), MD/JSON reports, SPA methodology, robust RL

2026-07-30 20:06:05 -03:00

hygiene.rs

feat(3.6.6): local/uncensored llama.cpp provider, clippy clean, CI (#40)

2026-08-03 23:48:08 -03:00

inbox.rs

feat(net): fail-closed egress, self-hosted OOB channel, inbound SMS

2026-09-14 01:35:36 -03:00

integrations.rs

feat: PR security gate, @neurosploit bot, richer NL REPL (#39)

2026-08-02 19:12:18 -03:00

integrity.rs

feat(hardening): scope-evasion resistance, evidence integrity, untrusted tool output

2026-09-18 21:52:14 -03:00

internal.rs

feat(internal): internal network / AD attack graph with choke-point analysis

2026-09-14 01:19:15 -03:00

knowledge_graph.rs

feat: attack knowledge graph, layered memory, command rectification, FAIR dashboard

2026-09-07 15:35:36 -03:00

lib.rs

bench: refresh the TypeSafe benchmark — 13/13 coverage, data-type-aware severity

2026-09-20 14:06:39 -03:00

memory.rs

feat: attack knowledge graph, layered memory, command rectification, FAIR dashboard

2026-09-07 15:35:36 -03:00

models.rs

fix(models): treat subscription session/usage-limit stdout as exhaustion

2026-09-19 23:16:56 -03:00

netguard.rs

feat(hardening): scope-evasion resistance, evidence integrity, untrusted tool output

2026-09-18 21:52:14 -03:00

oob.rs

feat(net): fail-closed egress, self-hosted OOB channel, inbound SMS

2026-09-14 01:35:36 -03:00

pipeline.rs

feat(chain,skills): close benchmark misses — CRLF-on-Location, second-order precondition; condense BENCHMARK

2026-09-20 12:42:36 -03:00

poc.rs

fix(wiring): activate cvss.rs + waf.rs (were shelf-ware); TypeSafe into CVSS + agent selection

2026-09-19 18:06:50 -03:00

policy.rs

feat(harness): risk model, engagement policies, capability tokens, audit trail

2026-09-13 17:56:31 -03:00

pomdp.rs

v3.6.5: LLM red-teaming (jailbreaks & prompt injection) + Opus 5 / Sonnet 5 / Kimi K3

2026-07-28 13:38:15 -03:00

pool.rs

fix(models): treat subscription session/usage-limit stdout as exhaustion

2026-09-19 23:16:56 -03:00

probe.rs

feat(hardening): scope-evasion resistance, evidence integrity, untrusted tool output

2026-09-18 21:52:14 -03:00

prosecutor.rs

feat(prosecutor): a second judge that shrinks claims instead of deleting findings

2026-09-14 00:35:38 -03:00

provenance.rs

feat(budget,provenance): reasoning budget modes and JOASNSCOPE provenance

2026-09-14 01:14:03 -03:00

proxy.rs

feat: PoC validator, Kali sandbox, intercept proxy, compliance, +8 validators

2026-09-14 10:26:06 -03:00

replay.rs

feat: PoC validator, Kali sandbox, intercept proxy, compliance, +8 validators

2026-09-14 10:26:06 -03:00

report.rs

release: v4.1.0 — assurance layer, TypeSafe, hardening + benchmark

2026-09-20 12:25:23 -03:00

rl.rs

v3.4.x: intelligent agent selection, whitebox, recon/code agents, Gemini, artifacts, RL, XBOW GUI

2026-06-23 11:39:56 -03:00

sandbox.rs

feat: PoC validator, Kali sandbox, intercept proxy, compliance, +8 validators

2026-09-14 10:26:06 -03:00

scope.rs

feat(hardening): scope-evasion resistance, evidence integrity, untrusted tool output

2026-09-18 21:52:14 -03:00

taint.rs

feat(hardening): scope-evasion resistance, evidence integrity, untrusted tool output

2026-09-18 21:52:14 -03:00

transport.rs

fix(transport): exempt loopback from the fail-closed gate (unambiguous host)

2026-09-19 18:24:34 -03:00

types.rs

feat: PoC validator, Kali sandbox, intercept proxy, compliance, +8 validators

2026-09-14 10:26:06 -03:00

typesafe_agent.rs

feat(typesafe): confirmation-loop agent + --typesafe flag (on/off/auto) for A/B

2026-09-19 18:15:10 -03:00

typesafe.rs

feat(cvss,typesafe): data-type-aware impact + evidence back-fill; LinkedIn article

2026-09-20 12:34:34 -03:00

uncertainty.rs

feat(budget,provenance): reasoning budget modes and JOASNSCOPE provenance

2026-09-14 01:14:03 -03:00

validation.rs

feat: PoC validator, Kali sandbox, intercept proxy, compliance, +8 validators

2026-09-14 10:26:06 -03:00

waf.rs

feat(waf): tell the edge apart from the application

2026-09-14 01:39:20 -03:00