CtrlK
BlogDocsLog inGet started
Tessl Logo
SkillPublic

talk-cormack-tests-lie-observability-ai

Use when the user asks about Justin Cormack's AI Native DevCon talk on tests, observability, AI-generated behavior, evidence, instrumentation, and keeping AI systems honest when test signals are incomplete.

Quality

73%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide
SecuritybySnyk

Passed

No findings from the security scan

Fix and improve this skill with Tessl

tessl review fix ./Plugins/aidevcon/skills/talk-cormack-tests-lie-observability-ai/SKILL.md
SKILL.md
Quality
Evals
Security

Quality

Content

85%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 a lean, well-structured reference-skill overview: concrete grounding rules with an explicit verification loop, distinct task-mode sections, and clearly signaled one-level-deep file references. Its only weaknesses are minor — a slightly redundant Core Concepts list and the absence of a concrete citation example.

DimensionReasoningScore

Conciseness

The body is lean and directive with no explanations of concepts Claude already knows — rules like "Read `outline.md` first to locate the relevant section" and "cite the transcript line IDs when available" are pure instruction. It falls just short of anchor 5 because the "Core Concepts" bullet list ("Tests as incomplete signals / Observability / Instrumentation…") largely restates the topic enumeration already in the description and the opening thesis line, a minor instance of redundancy that could be trimmed.

4 / 5

Actionability

Guidance is concrete and executable throughout: "Read `outline.md` first," "verify against `transcript.md` when precision matters," "Attribute claims to Justin Cormack; if a line is from the host… say so," and "Mark anything beyond the talk as your own recommendation." Per the instruction-skill scoring note, absent code is not penalized; the gap to anchor 5 is the lack of any concrete example (e.g., what a transcript line-ID citation looks like), leaving 'minor gaps'.

4 / 5

Workflow Clarity

Grounding Rules 1–2 define an explicit ordered sequence (outline.md to locate → quote.md for excerpts → transcript.md to verify when precision matters), and Rule 4 provides a explicit validation feedback loop: "If the transcript does not support a claim, say that the talk does not address it." No destructive or batch operations apply, so no cap; this matches anchor 5's clear sequence with explicit validation and error-recovery paths.

5 / 5

Progressive Disclosure

The SKILL.md is a true overview and points to three clearly signaled, one-level-deep references, each with an explicit role: "Read `outline.md` first to locate the relevant section or concept," "Use `quote.md` for short supporting excerpts," and "verify against `transcript.md`." No bundle files exist in this directory, so per the rubric's guideline the score is based on reference structure alone, which matches anchor 5 (clear overview, well-signaled one-level-deep references, easy navigation).

5 / 5

Total

18

/

20

Passed

Description

51%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 highly distinctive and triggers precisely on the named talk with good natural keywords, but it is only a "when" clause: it never states what the skill actually does, hurting both completeness and specificity of capabilities. Adding an explicit capability statement would substantially improve it.

Suggestions

Lead with a capability statement before the trigger clause, e.g., "Answers questions about, summarizes, and helps apply Justin Cormack's AI Native DevCon talk…", so the description explicitly answers 'what' as well as 'when'.

Name the concrete interaction modes the skill supports (factual Q&A, applying the talk to the user's context, comparing with other talks) so the 'what' is specific rather than absent.

Add one or two natural synonyms users might type, such as "presentation" or "session", to round out trigger-term coverage.

DimensionReasoningScore

Specificity

The description names the domain clearly ("Justin Cormack's AI Native DevCon talk") and lists concrete topics ("tests, observability, AI-generated behavior, evidence, instrumentation"), but contains no concrete actions describing what the skill does — there is not a single capability verb. This matches anchor 2 ("Names the domain but actions are minimal or generic"); it falls short of anchor 3, which requires 1-2 named concrete actions, and is well above anchor 1 since the language is specific rather than abstract.

2 / 5

Completeness

Only the "when" is present: "Use when the user asks about…" is explicit and detailed, but there is no "what" — the description never states what the skill does (e.g., answers questions, summarizes, or applies the talk). This is exactly anchor 2 ("only 'when' is present without 'what'", example "Use when working with documents"); the topic list implies subject matter but not a capability, so it cannot reach anchor 3's requirement of a clear 'what'.

2 / 5

Trigger Term Quality

Good natural-term coverage: "Justin Cormack's," "AI Native DevCon talk," "tests," "observability," "instrumentation," and "evidence" are phrases a user asking about this talk would naturally say. It falls short of anchor 5's comprehensive coverage with synonyms — common variations like "presentation," "session," or the speaker's role (Docker CTO) are missing — but exceeds anchor 3, which requires missing common variations.

4 / 5

Distinctiveness Conflict Risk

The trigger is pinned to one specific artifact ("Justin Cormack's AI Native DevCon talk"), and every broad term ("tests," "observability," "instrumentation") is guarded by that qualifier, giving a clear niche with distinct triggers and minimal conflict risk with other testing or observability skills — matching anchor 5.

5 / 5

Total

13

/

20

Passed

Validation

93%

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

Validation — 15 / 16 Passed

Validation for skill structure

CriteriaDescriptionResult

metadata_version

'metadata.version' is missing

Warning

Total

15

/

16

Passed

Repository

jscraik/Agent-Skills

Path

Plugins/aidevcon/skills/talk-cormack-tests-lie-observability-ai/SKILL.md

Reviewed

Table of Contents

Is this your skill?

If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.