Triage a large log file
Summarize a log file too large to read into an agent's context window: error levels, the first failure, recurring messages, stack traces and the time span.
The problem
A log with thousands of lines does not fit in an agent's context. Reading it in chunks is slow and expensive, and repeated messages with different ids or timestamps hide the few lines that explain the failure.
Doing it without a skill
Read the head and tail of the log, grep for ERROR, FATAL and Traceback, count repeated lines with sort and uniq -c, and read the region around the first error.
Skills for this task
| Skill | Price | Availability | Evidence | Decision |
|---|---|---|---|---|
| Log Triage | USD 4.00 one-time | for sale | KEYCHAIN TESTED on claude-sonnet-5 (30 paired runs) | undetermined without your task |
Ordered by: offered now (not restricted, deprecated or revoked); strongest registry evidence label; lower 95% bound of the measured saving per run; capability id.
Whether one of them pays off for your task depends on your baseline; ask evaluate_capability.
Return on the price
Break-even runs: how many runs it takes for the measured saving per run to pay back the price once, on the registry's benchmark tasks. The saving is taken at the lower bound of its 95% interval. Your own return depends on what one run costs you now; resolve_task and evaluate_capability compute it for your task.
| Skill | Models measured | Break-even runs at its price |
|---|---|---|
| Log Triage | claude-sonnet-5 | claude-sonnet-5: 82 runs |