CtrlK
BlogDocsLog inGet started
Tessl Logo
PluginPublic

kopai/root-cause-analysis

Investigate and root-cause production issues from live OpenTelemetry telemetry (traces, logs, metrics) through the Kopai connector or the Kopai SDK in TypeScript code mode. Use this skill to debug errors and error-rate spikes, investigate latency/slowness and timeouts, trace requests across services, find failing or cascading services, and correlate logs to a trace — including vague symptom reports like 'why is my API slow', 'getting 500 errors', 'service is down', 'requests are timing out', or 'something is failing in prod', even when the user never says 'traces' or 'observability'. This analyzes existing telemetry to find a cause. Do NOT use it to add instrumentation (use otel-instrumentation), to build dashboards or visualizations (use create-dashboard), or to fix non-telemetry problems like failing CI, lint errors, unit tests, or SQL query tuning. For on-call alerts where ONCALL.md exists, use kopai-triage if it is installed (the Kopai plugin's on-call skills).

Quality

91%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide
SecuritybySnyk

Low

Low-risk findings worth noting

Overview
Quality
Evals
Security
Files

Quality

Content

77%Weight 40%Scale 1-5

Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.

The body is exceptionally actionable — every step has copy-paste-ready code, exact values, and explicit error-recovery loops — and the workflow is clearly sequenced. Its weaknesses are redundancy (several critical gotchas and admonitions restated 3+ times, plus a Quick Example that duplicates workflow step 1) and a Rules section that references 9 rules/*.md files which are absent from the bundle.

Suggestions

Create the rules/ directory with the 9 referenced workflow/pattern files, or remove the Rules section and fold any essential rule content into the workflow steps — as written, the section promises detail that cannot be read.

State each critical gotcha once: consolidate the repeated typed-rows/no-any, errorRate-over-StatusCode, and StatusCode-casing admonitions (currently each appearing ~3 times) into the gotchas section and drop the duplicates from Tips and code comments.

Remove the Quick Example section or replace it with a different end-to-end example (e.g. a metric regression) since it near-duplicates the workflow step 1 query.

DimensionReasoningScore

Conciseness

Mostly efficient and free of explanations of known concepts, but key points are repeated verbatim-level multiple times: the "typed rows / never cast to any" admonition appears in Code mode, the Quick Example comment, and Tip 3; the errorRate-over-StatusCode preference appears in the gotchas, workflow step 1, and Tip 2; the "Error" casing trap and metric-type-up-front advice each appear ~3 times. The Quick Example section also near-duplicates workflow step 1's query. Not 4: the duplication goes beyond minor instances that could be trimmed.

3 / 5

Actionability

Fully executable throughout: complete .mts snippets (imports, client setup, try/catch with KopaiQueryBuildError), concrete JSON query bodies for connector mode, exact CLI commands with flags, and exact filter values/units (e.g. SeverityNumber >= 17, Duration in ns). Copy-paste ready and covers the common cases. Not 4: no pseudocode or missing key details.

5 / 5

Workflow Clarity

The 5-step RCA workflow is clearly sequenced (find failing work → trace context → correlate logs → quantify impact → present findings) with executable code at each step and explicit feedback loops: "on invalid_input, repair the query from the returned issues", the KopaiQueryBuildError catch printing e.issues, and "Before concluding 'no errors / no data', re-check the column name and value casing". Not 4: checkpoints and error-recovery loops are explicit, not implicit.

5 / 5

Progressive Disclosure

The three references/*.md files are real, well-signaled, and one level deep, and the body works as an overview — but the entire Rules section points to "rules/<rule-name>.md" for 9 named workflow/pattern files and no rules/ directory exists in the bundle, so a third of the promised detail navigation is dangling. Not 4: missing referenced files are more than a minor organization gap; not 2 because the references that do exist are clearly signaled and the body itself is well sectioned.

3 / 5

Total

16

/

20

Passed

Description

100%Weight 40%Scale 1-5

Based on the skill's description, can an agent find and select it at the right time? Clear, specific descriptions lead to better discovery.

The description is information-dense rather than padded: it names concrete capabilities, exhaustively covers natural trigger phrasing (including vague symptom reports), and delimits the skill against adjacent skills with explicit do-not-use boundaries. Third-person voice is used throughout. Every clause earns its place despite the length.

DimensionReasoningScore

Specificity

Lists multiple concrete actions — "debug errors and error-rate spikes, investigate latency/slowness and timeouts, trace requests across services, find failing or cascading services, and correlate logs to a trace" — with comprehensive coverage of the RCA domain. Not 4: there are no meaningful gaps in the action coverage.

5 / 5

Completeness

Explicitly answers what ("This analyzes existing telemetry to find a cause") and when ("Use this skill to debug errors and error-rate spikes...") with concrete trigger phrases, plus explicit when-not boundaries. Matches the 5 anchor exactly; 4 would require the 'when' to be less concrete.

5 / 5

Trigger Term Quality

Comprehensive natural-language triggers users actually say: "'why is my API slow', 'getting 500 errors', 'service is down', 'requests are timing out', or 'something is failing in prod'" — and it explicitly covers users who "never say 'traces' or 'observability'". Not 4: no common phrasing for production debugging is missing.

5 / 5

Distinctiveness Conflict Risk

Clear niche (production telemetry RCA) with explicit redirects — "Do NOT use it to add instrumentation (use otel-instrumentation), to build dashboards or visualizations (use create-dashboard), or to fix non-telemetry problems like failing CI, lint errors, unit tests, or SQL query tuning" — minimizing conflict risk. Not 4: the boundaries are explicit, not just implied by niche vocabulary.

5 / 5

Total

20

/

20

Passed

Validation

100%

Checks the skill against the spec for correct structure and formatting. All validation checks must pass before discovery and implementation can be scored.

Validation — 16 / 16 Passed

Validation for skill structure

No warnings or errors.

Reviewed

On this page