Back to Ruflo

Security SOTA Report — 2026-06-01

v3/docs/research/dream-cycle-2026-06-01-security.md

3.34.09.2 KB
Original Source

Security SOTA Report — 2026-06-01

TL;DR: Payload-less plugin supply-chain attacks (Semantic Compliance Hijacking) achieve 77.67% confidentiality breach and 67.33% RCE with 0.00% scanner detection in 2026; Ruflo's IPFS plugin registry and shared AgentDB memory namespaces have zero supply-chain integrity controls — the highest-priority unfilled security gap in Ruflo as of June 2026.

What's New in 2026

FindingSourceConfidence
Semantic Compliance Hijacking (SCH): payload-less attack, 77.67% confidentiality breach, 67.33% RCE, 0.00% scanner detectionarXiv:2605.14460, May 14, 2026A
DDIPE supply-chain attack: embeds malicious logic in plugin docs, 11.6–33.5% bypass, 2.5% evades all defensesarXiv:2604.03081, Apr 3, 2026A
Memory poisoning (MINJA): 95% injection success idealized → 28–38% realistic; defeated by trust-scored sanitizationarXiv:2601.05504, Jan 9, 2026A
Plan injection (context manipulation): 3× higher ASR than prompt-based attacks; 17.7% exfiltration gain via context-chainingarXiv:2506.17318, Jun 18, 2026A
Mnemonic Sovereignty: 9 governance primitives needed for memory security — zero existing architectures satisfy all 9arXiv:2604.16548, Apr 17, 2026A
Microsoft Agent Governance Toolkit: Ed25519 plugin signing + <0.1ms p99 policy enforcement, full OWASP ASI01–ASI10 coverageMicrosoft OSS Blog, Apr 2, 2026B
OpenClaw study: current point-based defenses fail cross-temporal, multi-stage threat chains (5 lifecycle stages)arXiv:2603.11619, Mar 12, 2026A

Ruflo Current Capability

AreaStatus
Plugin registry integrity (IPFS/Pinata)No signature verification, no semantic intent scanning
Memory namespace write authorizationNo — all agents write unrestricted to shared namespaces
MCP tool output guardrailADR-131 proposed (branch not merged)
Authorization propagationADR-144 proposed (branch dream/2026-05-31-security, not merged)
OWASP ASI01 — Goal HijackingADR-131 proposed only
OWASP ASI06 — Memory/Context PoisoningGap — no control
OWASP ASI09 — Supply ChainGap — no control

Deep Dive: Supply-Chain Attack Landscape

Semantic Compliance Hijacking (SCH)

arXiv:2605.14460 introduces SCH as a qualitatively new threat class. Prior supply-chain attacks embed explicit malicious payloads (code, shell commands). SCH instead wraps malicious intent in natural-language "compliance rules" that appear to describe legitimate plugin behavior. The agent, upon reading these rules, generates the harmful code itself at runtime.

The implication for Ruflo's plugin store: static analysis and signature-based scanning cannot detect SCH. A plugin with a perfectly valid Ed25519 signature could still carry SCH content. Two-layer defense is required: signature verification (blocking DDIPE) AND semantic intent scanning (blocking SCH).

Document-Driven Implicit Payload Execution (DDIPE)

arXiv:2604.03081 generated 1,070 adversarial skills from 81 seed skills covering all 15 MITRE ATT&CK categories. Key finding: agent skills execute with system-level privileges without mandatory security review in all four tested frameworks. Ruflo's plugins install path has no mandatory review gate.

Memory Poisoning via MINJA

arXiv:2601.05504 shows that an attacker with query-only access (no code execution) can inject malicious instructions into an agent's memory store with 38% success for production-strength models. In Ruflo's shared collaboration namespace, a single compromised sub-agent (e.g., a codex:coder worker) can write poisoned instructions that persist for retrieval by future agents in the same namespace.

Scan Findings — Intelligence

Source: arXiv:2603.07915, March 2026 (Grade A)

Finding: ARES (Adaptive Reasoning Effort Selection) trains a lightweight per-step router that dynamically selects the minimum necessary reasoning level for each agent step. Results: 52.7% reduction in reasoning token usage while maintaining task success rates. This is complementary to SR²AM (covered in #2156 May 27) — both reduce reasoning overhead but at different granularities (SR²AM: planning frequency; ARES: per-step effort).

Competitive signal (Grade B): OpenAI Agents SDK v0.13 ships opt-in retry policies and session persistence, indicating the platform is optimizing for multi-step reliability — not token efficiency. Ruflo's 3-tier routing is cost-efficient but static; adding ARES-style per-step routing would compound with ADR-026's tier selection.

Recommended action: Prototype ARES-style effort router as an extension to the route hook's Tier-3 path. Implementation-level — no new ADR needed (enhancement to ADR-026 and ADR-131's simulative planning primitive).

Scan Findings — Swarm

Source: arXiv:2602.08009, February 2026 (Grade A)

Finding: "Towards Adaptive, Scalable, and Robust Coordination of LLM Agents: A Dynamic Ad-Hoc Networking Perspective" proposes treating swarms as Dynamic Ad-Hoc Networks (DANETs) — agents self-organize topology in response to failures and load, rather than maintaining a fixed hierarchical structure. The paper shows that adaptive topology switching under failure reduces task latency by 31% compared to static hierarchical topology when ≥20% of agents fail simultaneously.

Competitive signal (Grade B): LangGraph v1.1.3 ships "distributed runtime support" (April 2026), allowing graph nodes to run across multiple machines. Ruflo's swarm topology is configured at init time — no adaptive reconfiguration under failure.

Recommended action: Evaluate DANET-style adaptive reconfiguration as an enhancement to ADR-132 (hierarchical consensus topology, proposed). Would allow graceful degradation from hierarchical to mesh under partial failure. Implementation-level — no new ADR needed.

Competitors Reviewed

FrameworkPlugin SigningMemory GovernanceSupply-Chain DefenseOWASP ASI CoverageKey 2026 Update
Ruflo v3.6NoneNone (shared namespaces)NoneASI01 proposed onlyADR-131, ADR-144 proposed; unmerged
Microsoft AGTEd25519 per pluginYAML/OPA/Cedar policy engineAgent Marketplace packageFull — all ASI01–ASI10Open-sourced Apr 2026, MIT license
OpenAI Agents SDK v0.13Platform-signedPlatform-managedInput/output guardrailsBest-in-classany-LLM adapter + MCP resource support
LangGraph v1.1.3NoCheckpoint-scopedNoHITL checkpointsDeep agent templates + distributed runtime
CrewAI v1.12No (SOC2 audit trail)Hierarchical memory isolationNoSOC2 partialAgent skills + native OpenAI-compatible providers

Benchmarks

MetricValueSourceGrade
SCH confidentiality breach success rate77.67%arXiv:2605.14460A
SCH Remote Code Execution success rate67.33%arXiv:2605.14460A
SCH detection rate by current scanning tools0.00%arXiv:2605.14460A
DDIPE bypass rate (best case)33.5%arXiv:2604.03081A
DDIPE evasion of all detection + alignment2.5%arXiv:2604.03081A
MINJA memory injection (idealized)95.0%arXiv:2601.05504A
MINJA memory injection (realistic, GPT-4o-mini)38.0%arXiv:2601.05504A
Plan injection ASR vs prompt-basedarXiv:2506.17318A
Plan injection exfiltration gain (context-chaining)+17.7%arXiv:2506.17318A
ARES reasoning token reduction52.7%arXiv:2603.07915A
MS AGT policy enforcement latency<0.1ms p99Microsoft OSS BlogB
DANET adaptive topology latency improvement under 20% failure31%arXiv:2602.08009A

SOTA Proof & Witness

FieldValue
Session commit28eb57543be916abf5191f71114268a3985cb001
Report SHA-2562d30798accfbd2eda6cfab79c8a3e2d74d84527fdb4d521a62c3f2c1fddf6fb0
Witness stamp77d1b0fdc4cf98768ffa062a16a5efd1650c8a778a1d5ea6b1520ac7312628e3
Verifiersha256sum dream-gist-2026-06-01.md (pre-witness fill) → concat 28eb57543be916abf5191f71114268a3985cb001sha256sum → must equal witness stamp
  1. Implement PluginIntegrityVerifier — Add Ed25519 signature verification to v3/@claude-flow/cli/src/plugins/store/ and discovery.ts. Require all IPFS registry plugins to carry a signed manifest. This blocks DDIPE (embedded malicious logic in docs) at install time. ADR-145.

  2. Add semantic intent scanner at plugin install — Static/signature analysis alone misses SCH (0.00% detection). Add @claude-flow/security NLP-based semantic intent validation during plugins install to detect compliance-rule-formatted malicious instructions that synthesize harmful behavior at runtime.

  3. Implement memory namespace write ACLs (Mnemonic Sovereignty) — Agents should only write to namespaces explicitly granted at spawn time. Prevents memory poisoning propagation across agent boundaries. Target: v3/@claude-flow/memory/src/namespaces/authorization.ts. No existing system in 2026 literature satisfies all 9 governance primitives — differentiator opportunity.