skills.rest
Zhachory1
Zhachory1COMMUNITY
Explore all skills by this author →
Website𝕏@zhachory1GitHub

run-telemetry

Emit structured JSONL telemetry events for AI agent workflow runs.

Updated Jul 8, 2026
Search Tags:#cli#observability#telemetry#schema validation#jsonl#agent workflows#run metrics
One-click install
npx skills add https://github.com/Zhachory1/autopraxis --skill run-telemetry-zhachory1
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: run-telemetry
Source: https://github.com/Zhachory1/autopraxis/tree/main/skills/run-telemetry
Command: npx skills add https://github.com/Zhachory1/autopraxis --skill run-telemetry-zhachory1

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve? AI agent workflows produce no durable evidence of what happened during a run, making it impossible to measure latency, cost, failure rates, or loop behavior and turn optimization into more than anecdotes. This Skill standardizes how agents record run metrics so downstream analysis can compare workflows and drive improvements. ## Core Features & Use Cases - Structured Event Emission: Append schema-validated JSONL events for start, end, gate, loop, escalation, validation, and human_response boundaries to .workflow-runs/<run-id>/telemetry.jsonl. - CLI Tooling: Use autopraxis telemetry emit|validate|summarize to write events, check them against the v1 JSON schema, and aggregate latency, cost, loop, and failure summaries. - Lifecycle Signal Aggregation: Run autopraxis telemetry lifecycle to compute skill usage, unmet-need notes, and run disposition tallies across runs with a minimum-sample floor. - Use Case: After a plan-to-launch workflow finishes, summarize its telemetry file to see per-step latency, council verdicts, human-edit rate, and PR outcomes, then feed those metrics into a backprop improvement review. ## Quick Start Ask the agent to emit a telemetry event for the current workflow step using the run-telemetry schema and then summarize the run's telemetry file.

Frequently Asked Questions about run-telemetry

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I emit telemetry events for an AI agent workflow run?▼

Run `autopraxis telemetry emit --workflow <name> --step <name> --event <event> --status <status>` with optional `--run-id` and `--path` flags. The command appends one schema-validated JSONL event and creates parent directories automatically.

What metrics should agent workflow telemetry capture?▼

Capture workflow and step names, event type, status, latency, cost, token counts, model and provider, loop iteration and cap, human-edit rate, verdicts, and escalation reasons. Store pointers and summaries rather than raw logs or sensitive data.

Where are telemetry JSONL files stored by default?▼

Events default to `.workflow-runs/<run-id>/telemetry.jsonl` inside the repository. When no repo is available, the fallback path is `~/.autopraxis/runs/<run-id>/telemetry.jsonl`, and a caller-provided path overrides both.

How do I validate telemetry events against the schema?▼

Run `autopraxis telemetry validate --path <file>` to check events against the telemetry-event-v1 JSON schema. The validator also rejects sensitive-looking keys or values to prevent raw data leakage.

Why does the lifecycle command report insufficient_signal?▼

The lifecycle aggregator refuses to emit add or prune signals below the `--min-runs` threshold, which defaults to 5 runs. Aggregate more runs with `autopraxis telemetry lifecycle --path <file>` or lower the floor explicitly.

Related Skills

Explore More Software Engineering Skills→
sapatamuku-creatorsapatamuku-creator

observability-and-instrumentation

Implements structured logging, metrics, tracing, and alerting for production code.

Community
Intermediate
brucebanner010198-commitsbrucebanner010198-commits

prompt-cache

Cache stable dispatch prompt segments with configurable TTL and telemetry.

Community
Advanced
EmeaAppGbbEmeaAppGbb

human-gate

Records human approvals and feedback at predefined checkpoints during software development phases.

Official
Advanced
MapleTechLabsMapleTechLabs
1.6k

maple-rust-style

Configure OpenTelemetry tracing, logs, and metrics for Rust services via OTLP HTTP to Maple.

Official
Intermediate