This skill should be used when the user asks to "query Langfuse traces", "show sessions", "check LLM costs", "analyse token usage", "view observations", "get scores", "query metrics", or mentions Langfuse, traces, or LLM observability. Also triggers on requests to analyse API latency, debug LLM calls, or investigate model performance.
/plugin marketplace add tavva/ben-claude-plugins/plugin install langfuse@ben-claude-pluginsThis skill inherits all available tools. When active, it can use any tool Claude has access to.
references/cli-reference.mdlf)Command-line interface for querying the Langfuse LLM observability platform. Query traces, sessions, observations, scores, and metrics with flexible filtering and output formats.
lf traces list [OPTIONS] # List traces with filters
lf traces get <ID> # Get specific trace
lf sessions list [OPTIONS] # List sessions
lf sessions show <ID> # Show session details
lf observations list [OPTIONS] # List observations (spans/generations)
lf observations get <ID> # Get specific observation
lf scores list [OPTIONS] # List scores
lf scores get <ID> # Get specific score
lf metrics query [OPTIONS] # Query aggregated metrics
lf traces list --limit 20
# Today's traces
lf traces list --from "$(date -u +%Y-%m-%dT00:00:00Z)"
# Last 24 hours
lf traces list --from "$(date -u -v-1d +%Y-%m-%dT%H:%M:%SZ)"
# Specific date range
lf traces list --from 2024-01-15T00:00:00Z --to 2024-01-16T00:00:00Z
lf traces list --user-id user123
lf traces list --session-id sess456
lf traces list --name "chat-completion"
lf traces list --tags production --tags v2
# Total cost over time
lf metrics query --view traces --measure total-cost --aggregation sum --granularity day
# Cost by model
lf metrics query --view observations --measure total-cost --aggregation sum --dimensions model
# Average cost per trace
lf metrics query --view traces --measure total-cost --aggregation avg
# P95 latency
lf metrics query --view traces --measure latency --aggregation p95
# Latency by trace name
lf metrics query --view traces --measure latency --aggregation avg --dimensions traceName
# Latency trends
lf metrics query --view traces --measure latency --aggregation p50 --granularity hour
# Total tokens
lf metrics query --view observations --measure total-tokens --aggregation sum
# Tokens by model
lf metrics query --view observations --measure total-tokens --aggregation sum --dimensions model
# Input vs output tokens
lf metrics query --view observations --measure input-tokens --aggregation sum
lf metrics query --view observations --measure output-tokens --aggregation sum
# Get trace details
lf traces get tr-abc123
# See all observations in a trace
lf observations list --trace-id tr-abc123
# Check scores for a trace
lf scores list --trace-id tr-abc123
All list and query commands support output format selection:
lf traces list --format table # Default, human-readable
lf traces list --format json # Machine-readable, full details
lf traces list --format csv # Spreadsheet-compatible
lf traces list --format markdown # Documentation-friendly
Save to file:
lf traces list --format json --output traces.json
The CLI uses profile-based configuration. Credentials resolve in order:
--public-key, --secret-key, --host)LANGFUSE_PUBLIC_KEY, LANGFUSE_SECRET_KEY, LANGFUSE_HOST)~/.config/langfuse/config.yml)lf config setup
lf traces list --profile production
The metrics command provides aggregated analytics:
Required parameters:
--view: traces or observations--measure: What to measure--aggregation: How to aggregateMeasures:
count - Number of itemslatency - Duration in millisecondsinput-tokens, output-tokens, total-tokens - Token countsinput-cost, output-cost, total-cost - Cost in USDAggregations:
count - Total countsum - Total sumavg - Averagep50, p95, p99 - Percentileshistogram - Distribution bucketsDimensions (group by):
traceName, model, environment, version, releaseGranularity (time bucketing):
auto, minute, hour, day, week, monthFor complete CLI documentation including all options:
references/cli-reference.md - Full command reference with all flagsCreating algorithmic art using p5.js with seeded randomness and interactive parameter exploration. Use this when users request creating art using code, generative art, algorithmic art, flow fields, or particle systems. Create original algorithmic art rather than copying existing artists' work to avoid copyright violations.
Applies Anthropic's official brand colors and typography to any sort of artifact that may benefit from having Anthropic's look-and-feel. Use it when brand colors or style guidelines, visual formatting, or company design standards apply.
Create beautiful visual art in .png and .pdf documents using design philosophy. You should use this skill when the user asks to create a poster, piece of art, design, or other static piece. Create original visual designs, never copying existing artists' work to avoid copyright violations.