Score breakdown
Popularity is tracked separately. Support, ads, sponsorships, and tips never affect these signals.
Why it matters
Useful for platform teams shipping AI agent workloads that need endpoint visibility, for security teams that want a single event model across desktop / CLI / IDE / gateway agents with versioned NDJSON records and JSON Schemas, for anyone who needs forensic reconstruction from on-disk session artifacts without depending on prior instrumentation, and for anyone who wants a CEL-based detection layer
Who should use it
Who should skip it
Skip perplexityai/numbat unless the captured evidence suggests it solves a problem you are actively working on.
About this signal
perplexityai/numbat is tracked by RepoRadar as a developer tool in the AI Observability section. First seen 2026-08-17; the source record was last checked on 2026-08-17. The current verdict is 'try now' with a Silver tier and moderate setup difficulty. Across RepoRadar's eight signals, perplexityai/numbat is strongest on workflow potential (8.5) and open-source/build quality (8.4) and weakest on setup ease (6.0) — a profile worth weighing against your own priorities. This page summarizes the evidence RepoRadar captured from https://github.com/perplexityai/numbat.
How this item is evaluated
The perplexityai/numbat record combines a 7.4/10 composite score with separate popularity (100.0), risk (conditional), and setup (moderate) signals. See the scoring methodology for the current weights and evidence definitions.
Putting this into practice? Read How to evaluate an AI tool before you adopt it for the checklist behind this score.
Risk explanation
Opt-in blocking is limited to supported synchronous pre-action hooks -- the coverage matrix at docs/agent-coverage.md#matrix is the source of truth, so an unsupported agent path cannot be blocked even with `enforce: true` rules; Shipped rules are monitor-only by default; promoting any rule to blocking requires explicit `enforce: true`, so a ruleset that should block will not block until that flag is set per-rule; The detection layer assumes the underlying agent's session artifacts are in a parseable format -- an agent that encrypts or rotates its on-disk artifacts (e.g. random session-id filenames with no stable schema) will degrade the forensic-reconstruction accuracy; The project is new enough that the shipped rule pack is small (the coverage matrix is the current state); the custom YAML rule path is the production extension surface, so a real rollout needs an internal rule-writing workflow.