Help us improve
Share bugs, ideas, or general feedback.
From agent-capability-standard
Establishes cause-effect relationships between events or states for root cause analysis, dependency mapping, tracing effects, and building causal models.
npx claudepluginhub synaptiai/synapti-marketplace --plugin agent-capability-standardHow this skill is triggered — by the user, by Claude, or both
Slash command
/agent-capability-standard:attributeThis skill is limited to the following tools:
The summary Claude sees in its skill listing — used to decide when to auto-load this skill
Establish causal relationships between observed effects and potential causes. This capability supports root cause analysis, dependency mapping, and causal reasoning for planning and debugging.
Systematically investigates causal relationships to identify true root causes rather than correlations or symptoms. Tests competing explanations and designs interventions addressing underlying drivers.
Maps causal relationships, traces dependencies, and reasons about consequences. Use for root cause analysis, impact assessment, and dependency tracing.
Diagnoses causes of phenomena using abductive reasoning: separates observations from interpretations, applies lenses (technical, human-factors, process, etc.), evaluates candidates against evidence, calibrates confidence, reports leaders with distinguishing tests. Feedback only.
Share bugs, ideas, or general feedback.
Establish causal relationships between observed effects and potential causes. This capability supports root cause analysis, dependency mapping, and causal reasoning for planning and debugging.
Success criteria:
Compatible schemas:
schemas/output_schema.yaml| Parameter | Required | Type | Description |
|---|---|---|---|
effect | Yes | any | The observed effect to attribute |
candidates | No | array | Potential causes to evaluate |
context | No | object | Situational context for attribution |
depth | No | string | Analysis depth: immediate, chain, comprehensive |
Characterize the effect: Understand what needs to be explained
Identify candidate causes: Generate list of potential causes
Evaluate causal strength: Assess each candidate
Trace causal chain: Map the path from cause to effect
Rule out alternatives: Eliminate unlikely causes
Quantify confidence: Assess attribution certainty
Return a structured object:
causes:
- cause: string # Identified cause
strength: number # 0.0-1.0 causal strength
mechanism: string # How cause produces effect
evidence: array[string] # Supporting evidence
type: string # direct, contributing, enabling
strengths: # Comparative strengths
cause_1: number
cause_2: number
causal_chain:
- step: string # Intermediate step
from: string # Prior state/event
to: string # Resulting state/event
alternatives_ruled_out:
- cause: string
reason: string # Why ruled out
confidence: 0..1
evidence_anchors: ["file:line", "log:entry"]
assumptions: []
| Field | Type | Description |
|---|---|---|
causes | array | Identified causes with strength and mechanism |
causes[].strength | number | Causal strength (0-1) |
causes[].mechanism | string | Explanation of causal pathway |
causes[].type | string | Direct, contributing, or enabling cause |
causal_chain | array | Steps from cause to effect |
alternatives_ruled_out | array | Rejected alternatives with reasons |
Input:
effect: "API response time increased from 200ms to 2000ms"
candidates:
- "recent deployment"
- "database load increase"
- "memory leak"
context:
timing: "started 2024-01-15 14:30"
affected: "/api/users endpoint"
Output:
causes:
- cause: "N+1 query introduced in recent deployment"
strength: 0.85
mechanism: "New eager loading removed; each user now triggers 3 additional queries"
evidence:
- "Deployment at 14:25, issue started 14:30"
- "Query count per request increased from 2 to 50+"
- "Database CPU correlates with API latency"
type: "direct"
- cause: "Database connection pool saturation"
strength: 0.40
mechanism: "Increased queries exhaust connection pool, causing wait times"
evidence:
- "Pool utilization at 95%"
- "Connection wait time increased"
type: "contributing"
strengths:
n_plus_1_query: 0.85
connection_pool: 0.40
memory_leak: 0.05
causal_chain:
- step: "Code deployed"
from: "eager_load(:profile)"
to: "no eager loading"
- step: "Query multiplication"
from: "2 queries/request"
to: "50+ queries/request"
- step: "Database overload"
from: "normal CPU"
to: "90% CPU"
- step: "Response delay"
from: "200ms"
to: "2000ms"
alternatives_ruled_out:
- cause: "memory leak"
reason: "Memory usage stable, no growth pattern"
- cause: "external service latency"
reason: "No external calls in affected endpoint"
confidence: 0.85
evidence_anchors:
- "git log --since='2024-01-15 14:00'"
- "app/models/user.rb:45 (diff)"
- "monitoring:query_count"
- "monitoring:db_cpu"
assumptions:
- "Monitoring data is accurate"
- "No other deployments in window"
Input:
effect: "UserService#create test failing with 'Email already taken'"
candidates:
- "test data pollution"
- "code change"
- "database state"
Output:
causes:
- cause: "Missing database cleanup between tests"
strength: 0.90
mechanism: "Test creates user, doesn't clean up, subsequent test finds duplicate"
evidence:
- "Test passes in isolation"
- "Test fails when run after create_user_test"
- "No database_cleaner configuration"
type: "direct"
strengths:
test_pollution: 0.90
code_change: 0.05
fixture_issue: 0.05
causal_chain:
- step: "First test runs"
from: "empty database"
to: "user 'test@example.com' exists"
- step: "Second test attempts create"
from: "user exists"
to: "uniqueness validation fails"
alternatives_ruled_out:
- cause: "code change"
reason: "No recent changes to UserService"
- cause: "fixture issue"
reason: "Test doesn't use fixtures"
confidence: 0.90
evidence_anchors:
- "spec/services/user_service_spec.rb:45"
- "spec/rails_helper.rb (no database_cleaner)"
- "command:rspec spec/services/user_service_spec.rb:45 --order defined"
assumptions:
- "Test failure is deterministic with this order"
- "No parallel test execution"
Verification tools: Read, Grep (to verify evidence)
mutation: falserequires_checkpoint: falserequires_approval: falserisk: lowCapability-specific rules:
Commonly follows:
detect - Detect anomalies then attribute causesobserve - Observations reveal effects to attributemeasure - Measurements quantify effectsCommonly precedes:
plan - Causal understanding informs remediationexplain - Attribution enables explanationcritique - Causal model enables critiqueAnti-patterns:
predict)Workflow references:
reference/workflow_catalog.yaml#debug_code_change for attribution in debuggingreference/workflow_catalog.yaml#world_model_build for causal modeling