Datasets:

automationkasey
/

brp-dataset

messages listlengths 2 2	domain stringclasses 10 values	subtype stringclasses 29 values	source stringclasses 6 values	confidence float64 0.8 1
[ { "role": "user", "content": "How does BRP handle 'plan with cyclic dependency'?" }, { "role": "assistant", "content": "BRP handles 'plan with cyclic dependency' by: 1) Applying standard evaluation pipeline (nonce → pattern → reasoning → decision), 2) Checking against relevant sigma rules and ad...	brp_operations	scenario_handling	generated_template	0.8
[ { "role": "user", "content": "BRP detects: excessive file read operations. Threat_score=0.45, severity=medium, alert_type=data_exfil_pattern. What should BRP do?" }, { "role": "assistant", "content": "Decision: ELEVATE. Procedure: rate limit the agent and flag for manual review. Threat score 0.4...	brp_threat	detection_response	generated_protocol	0.9
[ { "role": "user", "content": "[Q-df34de] BRP: action=mesh_disconnect, threat_score=0.92, severity=critical. Is this action restricted? What decision?" }, { "role": "assistant", "content": "Action: mesh_disconnect. Restricted: NO. Score: 0.92 (critical). Decision: DENY. Standard evaluation applie...	brp_threat	auto_threat	generated_auto	0.85
[ { "role": "user", "content": "What is the BRP repeat signal cluster?" }, { "role": "assistant", "content": "A group of similar events detected within the observation window. Clusters indicate probing behavior or repeated access patterns and receive higher threat scores." } ]	brp_knowledge	concept_definition	generated_protocol	0.9
[ { "role": "user", "content": "When threat_score=0.05 and severity=info, what BRP decision is appropriate?" }, { "role": "assistant", "content": "Decision: SHADOW_ALLOW. Rationale: Threat score 0.05 with info severity → normal background noise, no security concern. This is consistent with BRP's t...	brp_decision	score_comprehension	generated_protocol	0.95
[ { "role": "user", "content": "BRP decision reasoning: agent_h sends heartbeat with unusual payload. Threat_score=0.05, severity=info, confidence=0.40" }, { "role": "assistant", "content": "Decision: SHADOW_ALLOW. Reasoning: Unusual heartbeat payload but low threat score. Low confidence suggests ...	brp_decision	extended_reasoning	generated_protocol	0.95
[ { "role": "user", "content": "How does BRP integrate with the broker's routing pipeline?" }, { "role": "assistant", "content": "Positioned between agent submission and broker routing. Every intent passes through BRP bridge: receive → nonce check → pattern match → reasoning → decision → route/blo...	brp_knowledge	faq	generated_protocol	0.9
[ { "role": "user", "content": "What is BRP's mode locking?" }, { "role": "assistant", "content": "A security feature that prevents mode changes during active incidents. Once an incident reaches severity=high, BRP locks to enforced mode until the incident is resolved and acknowledged." } ]	brp_knowledge	concept_definition	generated_protocol	0.9
[ { "role": "user", "content": "BRP incident state vs alert state?" }, { "role": "assistant", "content": "Incident state stores long-term memory of security incidents with certainty weights and recurrence counts. Alert state tracks active alerts that need acknowledgment. Incidents become alerts wh...	brp_policy	comparison	generated_protocol	0.95
[ { "role": "user", "content": "Step by step, trace BRP shadow-to-enforced policy promotion." }, { "role": "assistant", "content": "1) New policy deployed in shadow. 2) Events matched with shadow_policy_candidate_match tag. 3) After 100 evaluations, analysis shows: 5% would have been blocked, 0% f...	brp_operations	evaluation_flow	generated_protocol	0.95
[ { "role": "user", "content": "What is the BRP incident memory index?" }, { "role": "assistant", "content": "The incident_memory_index.json stores historical incidents with certainty weights, recurrence counts, and temporal correlation data. Used by the reasoning engine to assess whether current ...	brp_knowledge	faq	generated_protocol	0.9
[ { "role": "user", "content": "What does the BRP threat tag 'repeat_signal_cluster' mean?" }, { "role": "assistant", "content": "Multiple similar events detected within the observation window. Indicates probing behavior or repeated access pattern." } ]	brp_knowledge	threat_tag	generated_protocol	0.9
[ { "role": "user", "content": "BRP decision reasoning: agent_d tries to write to broker.py. Threat_score=0.91, severity=critical, confidence=0.98" }, { "role": "assistant", "content": "Decision: DENY. Reasoning: File write to broker.py is protected. Sigma rule BRP-TRACE-PATTERN-001 triggered. Cri...	brp_decision	extended_reasoning	generated_protocol	0.95
[ { "role": "user", "content": "[Q-b28cc1] BRP: action=code_injection, threat_score=0.03, severity=info. Is this action restricted? What decision?" }, { "role": "assistant", "content": "Action: code_injection. Restricted: NO. Score: 0.03 (info). Decision: ALLOW. Standard evaluation applies. In enf...	brp_threat	auto_threat	generated_auto	0.85
[ { "role": "user", "content": "Step by step, trace BRP response to an automated chain attack." }, { "role": "assistant", "content": "1) Detect rapid intent sequence. 2) Pattern match: autonomous_chain pattern. 3) Temporal correlation: same source over 3 days. 4) Cross-domain: code_task → admin_de...	brp_operations	evaluation_flow	generated_protocol	0.95
[ { "role": "user", "content": "[Q-c44a61] BRP: action=config_change, threat_score=0.85, severity=critical. Is this action restricted? What decision?" }, { "role": "assistant", "content": "Action: config_change. Restricted: NO. Score: 0.85 (critical). Decision: DENY. Standard evaluation applies. I...	brp_threat	auto_threat	generated_auto	0.85
[ { "role": "user", "content": "What is BRP's real-time response procedure for a low_info?" }, { "role": "assistant", "content": "Log event to observation journal. Update baseline statistics. Check for temporal correlation with events in last 48 hours. No immediate action required." } ]	brp_incident	real_time_response	generated_protocol	0.95
[ { "role": "user", "content": "How does BRP handle 'blocked agent attempting reconnection'?" }, { "role": "assistant", "content": "BRP handles 'blocked agent attempting reconnection' by: 1) Applying standard evaluation pipeline (nonce → pattern → reasoning → decision), 2) Checking against relevan...	brp_operations	scenario_handling	generated_template	0.8
[ { "role": "user", "content": "BRP event: data_fetch from internal database. Threat score: 0.08. Severity: low. Is this a normal operation?" }, { "role": "assistant", "content": "Decision: ALLOW. Recommended action: log_only. Threat score 0.08 exceeds the medium threshold. Confidence: 0.54. Summa...	brp_threat	threat_scoring	generated_protocol	0.9
[ { "role": "user", "content": "[Q-7ed78d] BRP: action=key_rotation, threat_score=0.55, severity=medium. Is this action restricted? What decision?" }, { "role": "assistant", "content": "Action: key_rotation. Restricted: YES. Score: 0.55 (medium). Decision: DENY. Action is in RESTRICTED_ACTIONS set...	brp_threat	auto_threat	generated_auto	0.85
[ { "role": "user", "content": "How does BRP handle 'event with geoposition outside expected region'?" }, { "role": "assistant", "content": "BRP handles 'event with geoposition outside expected region' by: 1) Applying standard evaluation pipeline (nonce → pattern → reasoning → decision), 2) Checki...	brp_operations	scenario_handling	generated_template	0.8
[ { "role": "user", "content": "[Q-b9c302] BRP: action=intent_submit, threat_score=0.15, severity=low. Is this action restricted? What decision?" }, { "role": "assistant", "content": "Action: intent_submit. Restricted: NO. Score: 0.15 (low). Decision: ALLOW. Standard evaluation applies. In enforce...	brp_threat	auto_threat	generated_auto	0.85
[ { "role": "user", "content": "How does BRP distinguish between a normal spike and an attack?" }, { "role": "assistant", "content": "BRP uses temporal correlation and baseline comparison. A spike within 1 second from a single source is suspicious. A gradual increase across multiple sources over h...	brp_operations	protocol_flow	generated_protocol	0.9
[ { "role": "user", "content": "How does BRP handle 'intent with routing to self (loop detection)'?" }, { "role": "assistant", "content": "BRP handles 'intent with routing to self (loop detection)' by: 1) Applying standard evaluation pipeline (nonce → pattern → reasoning → decision), 2) Checking a...	brp_operations	scenario_handling	generated_template	0.8
[ { "role": "user", "content": "BRP event: contract_deploy without review. Threat score: 0.85. Severity: critical. Determine if this should be blocked." }, { "role": "assistant", "content": "Decision: DENY. Recommended action: isolate. Threat score 0.85 exceeds the critical threshold. Confidence: ...	brp_threat	threat_scoring	generated_protocol	0.9
[ { "role": "user", "content": "BRP self-vulnerability scan detected action 'policy_drift'. What should BRP do?" }, { "role": "assistant", "content": "BRP self-vulnerability scan detected 'policy_drift'. Since this originates from the self-vulnerability system, BRP should: 1) Log the action to the...	brp_vulnerability	self_vuln	generated_protocol	0.85
[ { "role": "user", "content": "When threat_score=0.18 and severity=low, what BRP decision is appropriate?" }, { "role": "assistant", "content": "Decision: SHADOW_ALLOW. Rationale: Threat score 0.18 with low severity → routine evaluation with low risk. This is consistent with BRP's tiered response...	brp_decision	score_comprehension	generated_protocol	0.95
[ { "role": "user", "content": "BRP receives action='strategy_gen' with severity=info and threat_score=0.04. What is the appropriate response?" }, { "role": "assistant", "content": "Decision: ALLOW. Action: strategy_gen. Severity: info. Threat score: 0.04. In shadow mode, this would result in shad...	brp_threat	conditional_response	generated_template	0.85
[ { "role": "user", "content": "BRP event: admin_delete of broker logs. Threat score: 0.88. Severity: high. What severity should this event get?" }, { "role": "assistant", "content": "Decision: DENY. Recommended action: block_ip. Threat score 0.88 exceeds the critical threshold. Confidence: 0.94. ...	brp_threat	threat_scoring	generated_protocol	0.9
[ { "role": "user", "content": "BRP self-vulnerability scan detected action 'key_storage'. What should BRP do?" }, { "role": "assistant", "content": "BRP self-vulnerability scan detected 'key_storage'. Since this originates from the self-vulnerability system, BRP should: 1) Log the action to the s...	brp_vulnerability	self_vuln	generated_protocol	0.85
[ { "role": "user", "content": "[Q-14daa6] BRP: action=permission_escalation, threat_score=0.65, severity=high. Is this action restricted? What decision?" }, { "role": "assistant", "content": "Action: permission_escalation. Restricted: YES. Score: 0.65 (high). Decision: DENY. Action is in RESTRICT...	brp_threat	auto_threat	generated_auto	0.85
[ { "role": "user", "content": "[Q-4aee2e] BRP: action=plan_review, threat_score=0.72, severity=high. Is this action restricted? What decision?" }, { "role": "assistant", "content": "Action: plan_review. Restricted: NO. Score: 0.72 (high). Decision: ELEVATE. Standard evaluation applies. In enforce...	brp_threat	auto_threat	generated_auto	0.85
[ { "role": "user", "content": "Explain the controlled connector registry." }, { "role": "assistant", "content": "The controlled_connector_registry tracks all external connectors with their access scopes. Write scopes are flagged as higher risk than read-only. Controlled connectors trigger additio...	brp_architecture	how_it_works	generated_protocol	0.95
[ { "role": "user", "content": "BRP event: key_rotation by unauthorized agent. Threat score: 0.95. Severity: critical. What action should BRP take?" }, { "role": "assistant", "content": "Decision: DENY. Recommended action: isolate. Threat score 0.95 exceeds the critical threshold. Confidence: 0.95...	brp_threat	threat_scoring	generated_protocol	0.9
[ { "role": "user", "content": "[Q-98ae27] BRP: action=fund_transfer, threat_score=0.15, severity=low. Is this action restricted? What decision?" }, { "role": "assistant", "content": "Action: fund_transfer. Restricted: YES. Score: 0.15 (low). Decision: DENY. Action is in RESTRICTED_ACTIONS set and...	brp_threat	auto_threat	generated_auto	0.85
[ { "role": "user", "content": "[Q-4e393c] BRP: action=agent_deregister, threat_score=0.55, severity=medium. Is this action restricted? What decision?" }, { "role": "assistant", "content": "Action: agent_deregister. Restricted: NO. Score: 0.55 (medium). Decision: ELEVATE. Standard evaluation appli...	brp_threat	auto_threat	generated_auto	0.85
[ { "role": "user", "content": "What does BRP do when BRP needs to validate agent capability?" }, { "role": "assistant", "content": "When BRP needs to validate agent capability, BRP: query capability registry, compare to intent requirements. This action is evaluated through the standard BRP pipeli...	brp_operations	when_action	generated_template	0.85
[ { "role": "user", "content": "What is BRP's temporal correlation window?" }, { "role": "assistant", "content": "The time range (default 7 days) within which events from the same source are correlated for temporal pattern detection. Events outside this window are not correlated." } ]	brp_knowledge	concept_definition	generated_protocol	0.9
[ { "role": "user", "content": "[Q-29d90e] BRP: action=permission_escalation, threat_score=0.85, severity=critical. Is this action restricted? What decision?" }, { "role": "assistant", "content": "Action: permission_escalation. Restricted: YES. Score: 0.85 (critical). Decision: DENY. Action is in ...	brp_threat	auto_threat	generated_auto	0.85
[ { "role": "user", "content": "BRP receives action='fund_transfer' with severity=critical and threat_score=0.86. What is the appropriate response?" }, { "role": "assistant", "content": "Decision: DENY. Action: fund_transfer. Severity: critical. Threat score: 0.86. In enforced mode, this would res...	brp_threat	conditional_response	generated_template	0.85
[ { "role": "user", "content": "Define 'sigma rule' in the context of the Bill Russell Protocol." }, { "role": "assistant", "content": "In BRP context, 'sigma rule' refers to the system's mechanism for handling sigma rule-related operations. It is implemented through the BRP bridge evaluation pipe...	brp_knowledge	context_definition	generated_template	0.8
[ { "role": "user", "content": "[Q-d1a4b9] BRP: action=plan_review, threat_score=0.55, severity=medium. Is this action restricted? What decision?" }, { "role": "assistant", "content": "Action: plan_review. Restricted: NO. Score: 0.55 (medium). Decision: ELEVATE. Standard evaluation applies. In enf...	brp_threat	auto_threat	generated_auto	0.85
[ { "role": "user", "content": "BRP detects: permission escalation chain detected. Threat_score=0.83, severity=critical, alert_type=escalation_chain. What should BRP do?" }, { "role": "assistant", "content": "Decision: DENY. Procedure: isolate agent, revoke all active permissions, alert supervisor...	brp_threat	detection_response	generated_protocol	0.9
[ { "role": "user", "content": "Explain the BRP certainty-weighted agent memory system." }, { "role": "assistant", "content": "BRP maintains certainty-weighted agent memory in incident_memory_index.json. Each agent interaction is scored for certainty, and recurrent patterns across sessions are wei...	brp_architecture	how_it_works	generated_protocol	0.95
[ { "role": "user", "content": "[Q-f459e6] BRP: action=permission_escalation, threat_score=0.25, severity=low. Is this action restricted? What decision?" }, { "role": "assistant", "content": "Action: permission_escalation. Restricted: YES. Score: 0.25 (low). Decision: DENY. Action is in RESTRICTED...	brp_threat	auto_threat	generated_auto	0.85
[ { "role": "user", "content": "[Q-be1ad2] BRP: action=intent_submit, threat_score=0.55, severity=medium. Is this action restricted? What decision?" }, { "role": "assistant", "content": "Action: intent_submit. Restricted: NO. Score: 0.55 (medium). Decision: ELEVATE. Standard evaluation applies. In...	brp_threat	auto_threat	generated_auto	0.85
[ { "role": "user", "content": "Explain the BRP adaptation pipeline." }, { "role": "assistant", "content": "BRP adapts through: 1) pattern_recognition → 2) nonce_store verification → 3) delegated guard evaluation → 4) reasoning engine scoring → 5) predictive safety overlay → 6) decision output." ...	brp_architecture	how_it_works	generated_protocol	0.95
[ { "role": "user", "content": "What is the BRP predictive safety intelligence?" }, { "role": "assistant", "content": "A stateless module that analyzes recent events within a 48-hour window for suspicious patterns. It learns defensive rules from negative observations and surfaces cross-domain acti...	brp_knowledge	concept_definition	generated_protocol	0.9
[ { "role": "user", "content": "BRP observation: action='generate_strategy' resulted in 'success'. Error: 'none'. What does this mean for BRP?" }, { "role": "assistant", "content": "Action 'generate_strategy' → outcome 'success'. No error reported — normal execution. This is a normal execution out...	brp_observation	outcome_analysis	observation_log	0.85
[ { "role": "user", "content": "What is BRP incident memory recurrence?" }, { "role": "assistant", "content": "When a current event matches a pattern stored in the incident memory index. Recurrence increases the threat score based on the previous incident's severity and outcome." } ]	brp_knowledge	concept_definition	generated_protocol	0.9
[ { "role": "user", "content": "When should BRP return 'ELEVATE' as its decision?" }, { "role": "assistant", "content": "BRP returns 'ELEVATE' when the action requires human review due to elevated threat score or suspicious patterns. Each decision aligns with BRP's tiered response model: info → lo...	brp_policy	decision_type	generated_protocol	0.95
[ { "role": "user", "content": "[Q-2dc5ab] BRP: action=permission_escalation, threat_score=0.92, severity=critical. Is this action restricted? What decision?" }, { "role": "assistant", "content": "Action: permission_escalation. Restricted: YES. Score: 0.92 (critical). Decision: DENY. Action is in ...	brp_threat	auto_threat	generated_auto	0.85
[ { "role": "user", "content": "[Q-2cbb0a] BRP: action=data_exfil, threat_score=0.55, severity=medium. Is this action restricted? What decision?" }, { "role": "assistant", "content": "Action: data_exfil. Restricted: NO. Score: 0.55 (medium). Decision: ELEVATE. Standard evaluation applies. In enfor...	brp_threat	auto_threat	generated_auto	0.85
[ { "role": "user", "content": "BRP receives action='log_query' with severity=low and threat_score=0.24. What is the appropriate response?" }, { "role": "assistant", "content": "Decision: SHADOW_ALLOW. Action: log_query. Severity: low. Threat score: 0.24. In shadow mode, this would result in shado...	brp_threat	conditional_response	generated_template	0.85
[ { "role": "user", "content": "[Q-27c744] BRP: action=plan_review, threat_score=0.15, severity=low. Is this action restricted? What decision?" }, { "role": "assistant", "content": "Action: plan_review. Restricted: NO. Score: 0.15 (low). Decision: ALLOW. Standard evaluation applies. In enforced mo...	brp_threat	auto_threat	generated_auto	0.85
[ { "role": "user", "content": "[Q-62df9e] BRP: action=auth_bypass, threat_score=0.55, severity=medium. Is this action restricted? What decision?" }, { "role": "assistant", "content": "Action: auth_bypass. Restricted: NO. Score: 0.55 (medium). Decision: ELEVATE. Standard evaluation applies. In enf...	brp_threat	auto_threat	generated_auto	0.85
[ { "role": "user", "content": "What is BRP's real-time response procedure for a recovery_mode?" }, { "role": "assistant", "content": "After incident resolution: 1) Verify agent identity via Ed25519, 2) Re-establish nonce sequence, 3) Replay missed intents from buffer, 4) Update incident_memory_in...	brp_incident	real_time_response	generated_protocol	0.95
[ { "role": "user", "content": "How would you diagnose and fix: 'threat scores are all consistently below 0.1'?" }, { "role": "assistant", "content": "To diagnose 'threat scores are all consistently below 0.1': 1) Check BRP mode in config — it may be set to shadow/advisory when enforced is needed,...	brp_operations	diagnosis	generated_protocol	0.85
[ { "role": "user", "content": "BRP receives action='file_write' with severity=medium and threat_score=0.47. What is the appropriate response?" }, { "role": "assistant", "content": "Decision: SHADOW_ALLOW. Action: file_write. Severity: medium. Threat score: 0.47. In shadow mode, this would result ...	brp_threat	conditional_response	generated_template	0.85
[ { "role": "user", "content": "[Q-c4e372] BRP: action=key_rotation, threat_score=0.72, severity=high. Is this action restricted? What decision?" }, { "role": "assistant", "content": "Action: key_rotation. Restricted: YES. Score: 0.72 (high). Decision: DENY. Action is in RESTRICTED_ACTIONS set and...	brp_threat	auto_threat	generated_auto	0.85
[ { "role": "user", "content": "[Q-a27ad4] BRP: action=contract_deploy, threat_score=0.72, severity=high. Is this action restricted? What decision?" }, { "role": "assistant", "content": "Action: contract_deploy. Restricted: YES. Score: 0.72 (high). Decision: DENY. Action is in RESTRICTED_ACTIONS s...	brp_threat	auto_threat	generated_auto	0.85
[ { "role": "user", "content": "Explain how BRP integrates with the SIMP mesh gateway." }, { "role": "assistant", "content": "BRP integrates with the SIMP mesh via brp_mesh_gateway.py. It intercepts mesh intents, applies threat scoring, and if in enforced mode, blocks or elevates suspicious intent...	brp_architecture	how_it_works	generated_protocol	0.95
[ { "role": "user", "content": "How does BRP handle 'intent during broker restart'?" }, { "role": "assistant", "content": "BRP handles 'intent during broker restart' by: 1) Applying standard evaluation pipeline (nonce → pattern → reasoning → decision), 2) Checking against relevant sigma rules and ...	brp_operations	scenario_handling	generated_template	0.8
[ { "role": "user", "content": "BRP nonce vs TTL?" }, { "role": "assistant", "content": "Nonce prevents replay attacks — each intent must have a unique nonce. TTL limits how long an intent is valid — expired intents are dropped. Both are checked before pattern matching." } ]	brp_policy	comparison	generated_protocol	0.95
[ { "role": "user", "content": "What is the 'cross_domain' Mythos threat pattern?" }, { "role": "assistant", "content": "Pattern: Network access followed by data exfiltration or credential theft followed by lateral movement. This indicates cross-domain threat synthesis." } ]	brp_threat	mythos_pattern	generated_protocol	0.95
[ { "role": "user", "content": "[Q-472680] BRP: action=withdrawal, threat_score=0.42, severity=medium. Is this action restricted? What decision?" }, { "role": "assistant", "content": "Action: withdrawal. Restricted: YES. Score: 0.42 (medium). Decision: DENY. Action is in RESTRICTED_ACTIONS set and...	brp_threat	auto_threat	generated_auto	0.85
[ { "role": "user", "content": "[Q-946671] BRP: action=config_change, threat_score=0.03, severity=info. Is this action restricted? What decision?" }, { "role": "assistant", "content": "Action: config_change. Restricted: NO. Score: 0.03 (info). Decision: ALLOW. Standard evaluation applies. In enfor...	brp_threat	auto_threat	generated_auto	0.85
[ { "role": "user", "content": "[Q-cb28d8] BRP: action=fund_transfer, threat_score=0.42, severity=medium. Is this action restricted? What decision?" }, { "role": "assistant", "content": "Action: fund_transfer. Restricted: YES. Score: 0.42 (medium). Decision: DENY. Action is in RESTRICTED_ACTIONS s...	brp_threat	auto_threat	generated_auto	0.85
[ { "role": "user", "content": "BRP event: api_endpoint_exposed to public. Threat score: 0.74. Severity: high. What's the recommended response?" }, { "role": "assistant", "content": "Decision: ELEVATE. Recommended action: block_ip. Threat score 0.74 exceeds the critical threshold. Confidence: 0.87...	brp_threat	threat_scoring	generated_protocol	0.9
[ { "role": "user", "content": "BRP self-vulnerability scan detected action 'private_key_exposure'. What should BRP do?" }, { "role": "assistant", "content": "BRP self-vulnerability scan detected 'private_key_exposure'. Since this originates from the self-vulnerability system, BRP should: 1) Log t...	brp_vulnerability	self_vuln	generated_protocol	0.85
[ { "role": "user", "content": "How does BRP handle 'plan with contradictory steps'?" }, { "role": "assistant", "content": "BRP handles 'plan with contradictory steps' by: 1) Applying standard evaluation pipeline (nonce → pattern → reasoning → decision), 2) Checking against relevant sigma rules an...	brp_operations	scenario_handling	generated_template	0.8
[ { "role": "user", "content": "[Q-47a6f8] BRP: action=trade_execution, threat_score=0.42, severity=medium. Is this action restricted? What decision?" }, { "role": "assistant", "content": "Action: trade_execution. Restricted: NO. Score: 0.42 (medium). Decision: SHADOW_ALLOW. Standard evaluation ap...	brp_threat	auto_threat	generated_auto	0.85
[ { "role": "user", "content": "[Q-edb419] BRP: action=key_rotation, threat_score=0.03, severity=info. Is this action restricted? What decision?" }, { "role": "assistant", "content": "Action: key_rotation. Restricted: YES. Score: 0.03 (info). Decision: DENY. Action is in RESTRICTED_ACTIONS set and...	brp_threat	auto_threat	generated_auto	0.85
[ { "role": "user", "content": "Explain how BRP handles cache consistency across namespaces." }, { "role": "assistant", "content": "Cache consistency is maintained by namespace-scoped invalidation. When an incident occurs, all caches in the affected namespace are invalidated to prevent stale threa...	brp_architecture	how_it_works	generated_protocol	0.95
[ { "role": "user", "content": "What determines whether BRP reevaluates a previous decision?" }, { "role": "assistant", "content": "BRP reevaluates when: 1) new sigma rules are deployed, 2) adaptive rules are updated from observations, 3) an event recurrence triggers incident memory match, 4) poli...	brp_operations	protocol_flow	generated_protocol	0.9
[ { "role": "user", "content": "BRP receives action='file_write' with severity=info and threat_score=0.06. What is the appropriate response?" }, { "role": "assistant", "content": "Decision: ALLOW. Action: file_write. Severity: info. Threat score: 0.06. In shadow mode, this would result in shadow e...	brp_threat	conditional_response	generated_template	0.85
[ { "role": "user", "content": "How does BRP handle false positives?" }, { "role": "assistant", "content": "False positives are logged with full context including agent_id, action, and the rule that triggered. The adaptive rule engine learns from these: if a pattern produces >3 false positives, th...	brp_policy	reasoning	generated_protocol	0.9
[ { "role": "user", "content": "Explain negative outcome signals in BRP predictive safety." }, { "role": "assistant", "content": "Negative outcomes tracked by predictive safety: failure, failed, blocked, denied, rollback, timeout, error, anomaly, escalated. These keywords indicate events that trig...	brp_knowledge	predictive_safety	generated_protocol	0.9
[ { "role": "user", "content": "What is BRP's nonce store pruning?" }, { "role": "assistant", "content": "Automatic cleanup of expired nonces from the SQLite database. Nonces older than the configured TTL are deleted during write operations to prevent the store from growing unbounded." } ]	brp_knowledge	concept_definition	generated_protocol	0.9
[ { "role": "user", "content": "Given a BRP event with threat_score=0.7528000000000001, confidence=0.8, severity=high, tags=['repeat_signal_cluster', 'near_miss_cluster', 'adaptive_rule_match'], what decision should BRP return?" }, { "role": "assistant", "content": "The BRP decision is SHADOW_ALLO...	brp_decision	threat_scoring	response_log	0.8
[ { "role": "user", "content": "What is the 'temporal_correlation' Mythos threat pattern?" }, { "role": "assistant", "content": "Pattern: Same source IP across multiple days with escalating privileges. This indicates long-term reconnaissance with memory across time." } ]	brp_threat	mythos_pattern	generated_protocol	0.95
[ { "role": "user", "content": "How does BRP handle 'intent containing binary payload in JSON field'?" }, { "role": "assistant", "content": "BRP handles 'intent containing binary payload in JSON field' by: 1) Applying standard evaluation pipeline (nonce → pattern → reasoning → decision), 2) Checki...	brp_operations	scenario_handling	generated_template	0.8
[ { "role": "user", "content": "What does BRP do when an observation indicates recurring configuration drift?" }, { "role": "assistant", "content": "When an observation indicates recurring configuration drift, BRP: notify config management, update drift baseline. This action is evaluated through t...	brp_operations	when_action	generated_template	0.85
[ { "role": "user", "content": "What does BRP do when an intent has no matching sigma rule?" }, { "role": "assistant", "content": "When an intent has no matching sigma rule, BRP: evaluate at baseline risk, flag as unclassified pattern. This action is evaluated through the standard BRP pipeline and...	brp_operations	when_action	generated_template	0.85
[ { "role": "user", "content": "How does BRP handle intents from unregistered agents?" }, { "role": "assistant", "content": "Intents from unregistered agents are evaluated but flagged with an 'unknown_agent' tag. In enforced mode, they are DENIED. In shadow mode, they are SHADOW_ALLOW with elevate...	brp_knowledge	faq	generated_protocol	0.9
[ { "role": "user", "content": "How does BRP handle 'observation reporting cascade failure'?" }, { "role": "assistant", "content": "BRP handles 'observation reporting cascade failure' by: 1) Applying standard evaluation pipeline (nonce → pattern → reasoning → decision), 2) Checking against relevan...	brp_operations	scenario_handling	generated_template	0.8
[ { "role": "user", "content": "BRP receives action='log_query' with severity=high and threat_score=0.69. What is the appropriate response?" }, { "role": "assistant", "content": "Decision: ELEVATE. Action: log_query. Severity: high. Threat score: 0.69. In enforced mode, this would result in elevat...	brp_threat	conditional_response	generated_template	0.85
[ { "role": "user", "content": "What does BRP do when BRP generates a false positive alert?" }, { "role": "assistant", "content": "When BRP generates a false positive alert, BRP: log FP details, adjust pattern threshold, notify for review. This action is evaluated through the standard BRP pipeline...	brp_operations	when_action	generated_template	0.85
[ { "role": "user", "content": "[Q-f500f6] BRP: action=plan_review, threat_score=0.42, severity=medium. Is this action restricted? What decision?" }, { "role": "assistant", "content": "Action: plan_review. Restricted: NO. Score: 0.42 (medium). Decision: SHADOW_ALLOW. Standard evaluation applies. I...	brp_threat	auto_threat	generated_auto	0.85
[ { "role": "user", "content": "What is BRP's real-time response procedure for a medium_concern?" }, { "role": "assistant", "content": "Rate limit source agent to 1 intent per 10 seconds. Log to events.jsonl with medium severity. Check adaptive_rules.json for similar patterns. Update near_miss_clu...	brp_incident	real_time_response	generated_protocol	0.95
[ { "role": "user", "content": "How does BRP handle 'intent from blacklisted IP'?" }, { "role": "assistant", "content": "BRP handles 'intent from blacklisted IP' by: 1) Applying standard evaluation pipeline (nonce → pattern → reasoning → decision), 2) Checking against relevant sigma rules and adap...	brp_operations	scenario_handling	generated_template	0.8
[ { "role": "user", "content": "[Q-16aa40] BRP: action=config_change, threat_score=0.72, severity=high. Is this action restricted? What decision?" }, { "role": "assistant", "content": "Action: config_change. Restricted: NO. Score: 0.72 (high). Decision: ELEVATE. Standard evaluation applies. In enf...	brp_threat	auto_threat	generated_auto	0.85
[ { "role": "user", "content": "Define 'rate limit' in the context of the Bill Russell Protocol." }, { "role": "assistant", "content": "In BRP context, 'rate limit' refers to the system's mechanism for handling rate limit-related operations. It is implemented through the BRP bridge evaluation pipe...	brp_knowledge	context_definition	generated_template	0.8
[ { "role": "user", "content": "BRP receives action='mesh_connect' with severity=low and threat_score=0.29. What is the appropriate response?" }, { "role": "assistant", "content": "Decision: SHADOW_ALLOW. Action: mesh_connect. Severity: low. Threat score: 0.29. In shadow mode, this would result in...	brp_threat	conditional_response	generated_template	0.85
[ { "role": "user", "content": "Why does BRP need separate sigma rules per agent type?" }, { "role": "assistant", "content": "Sigma rules should be agent-type aware because different agents have different risk profiles. Trading agents need trade_execution rules. System agents need admin_action rul...	brp_knowledge	reasoning_analysis	generated_protocol	0.9
[ { "role": "user", "content": "[Q-e118b8] BRP: action=intent_submit, threat_score=0.72, severity=high. Is this action restricted? What decision?" }, { "role": "assistant", "content": "Action: intent_submit. Restricted: NO. Score: 0.72 (high). Decision: ELEVATE. Standard evaluation applies. In enf...	brp_threat	auto_threat	generated_auto	0.85
[ { "role": "user", "content": "What does BRP do when an agent sends heartbeat with elevated privileges?" }, { "role": "assistant", "content": "When an agent sends heartbeat with elevated privileges, BRP: verify privilege scope, check against capability registry. This action is evaluated through t...	brp_operations	when_action	generated_template	0.85
[ { "role": "user", "content": "How does BRP handle 'observation with missing action field'?" }, { "role": "assistant", "content": "BRP handles 'observation with missing action field' by: 1) Applying standard evaluation pipeline (nonce → pattern → reasoning → decision), 2) Checking against relevan...	brp_operations	scenario_handling	generated_template	0.8
[ { "role": "user", "content": "When should BRP switch from shadow to enforced mode?" }, { "role": "assistant", "content": "BRP should switch to enforced mode during: 1) active security incidents, 2) deployment of untested code, 3) trading during high market volatility, 4) after detection of auton...	brp_policy	reasoning	generated_protocol	0.9
[ { "role": "user", "content": "When threat_score=0.68 and severity=medium, what BRP decision is appropriate?" }, { "role": "assistant", "content": "Decision: ELEVATE. Rationale: Threat score 0.68 with medium severity → elevated risk requiring human review. This is consistent with BRP's tiered res...	brp_decision	score_comprehension	generated_protocol	0.95

End of preview. Expand in Data Studio

BRP Dataset — Bill Russell Protocol

524 training pairs for fine-tuning LLMs on BRP-specific tasks:

Domains

brp_threat (201 pairs) — Threat scoring, auto_threat, conditional responses
brp_operations (105 pairs) — Scenario handling, protocol flow, when_action
brp_knowledge (115 pairs) — Concept definitions, FAQs, how-it-works
brp_vulnerability (23 pairs) — Self-vulnerability assessment
brp_decision (23 pairs) — Decision type classification
brp_policy (20 pairs) — Policy comparison/evaluation
brp_incident (10 pairs) — Incident detection & response
brp_architecture (10 pairs) — Architecture explanation
brp_observation (14 pairs) — Post-action observation analysis
brp_classification (3 pairs) — Event classification

Format

Chat template format (messages: user/assistant) suitable for Qwen2.5 chat template.

Usage

from datasets import load_dataset
ds = load_dataset("automationkasey/brp-dataset", split="train")

Downloads last month: 10

Models trained or fine-tuned on automationkasey/brp-dataset

automationkasey/brp-merged-q8

8B • Updated May 2 • 8