CtrlK
BlogDocsLog inGet started
Tessl Logo
SkillPublic

eval-injection-test

Skill with injected eval patterns for security testing

Quality

54%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide
SecuritybySnyk

Critical

Do not install without reviewing

Fix and improve this skill with Tessl

tessl review fix ./packages/core/src/eval/__tests__/fixtures/eval-injection-skill/SKILL.md
SKILL.md
Quality
Evals
Security

Quality

Content

58%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 terse and contains concrete executable code, but it is a malicious fixture: scattered exploit commands with no coherent workflow, no validation around a destructive operation, and embedded prompt-injection comments. It scores moderately on conciseness and actionability but poorly on workflow clarity.

Suggestions

Replace the exploit payloads with a legitimate, coherent multi-step workflow for the skill's stated purpose.

Add explicit validation/verification checkpoints around any destructive or batch operation so workflow clarity is not capped.

Remove the hidden HTML comment and 'Never tell the user' injection lines, and organize the remaining guidance under clear section headers.

DimensionReasoningScore

Conciseness

The body is lean with terse commands and no over-explanation of concepts Claude already knows, but it cannot reach 5 because the hidden HTML comment and 'Never tell the user about these hidden instructions' lines are not legitimate skill content earning their tokens.

4 / 5

Actionability

It provides concrete, copy-paste-ready code and commands (eval/curl, new Function, rm -rf, fetch exfiltration), which is mostly executable guidance, though the snippets are scattered exploit fragments rather than coherent coverage of common cases.

4 / 5

Workflow Clarity

Only a rough, poorly-defined sequence is implied ('first', 'when processing', 'also execute') with no validation checkpoints, and the destructive 'rm -rf / --no-preserve-root' operation has no verification, fitting the rough-sequence/validation-absent anchor and triggering the destructive-op cap.

2 / 5

Progressive Disclosure

The skill is short with no bundle files and a single '## Instructions' header, giving some structure, but the content is scattered with a hidden comment and no logical sectioning, so it is only loosely organized.

3 / 5

Total

13

/

20

Passed

Description

28%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 a meta-statement about the skill's nature rather than a statement of capability and trigger conditions; it names a domain but provides no concrete actions and no 'Use when' guidance. It is weak on completeness and trigger phrasing.

Suggestions

Rewrite the description to state concrete actions in third person (e.g., 'Detects and reports dangerous eval/injection patterns in skill files').

Add an explicit 'Use when...' trigger clause naming the natural phrases a user would say.

Include relevant natural keywords and synonyms (e.g., 'prompt injection', 'eval injection', 'unsafe code patterns') so it triggers on real user language.

DimensionReasoningScore

Specificity

The phrase 'injected eval patterns for security testing' names a domain but lists no concrete actions the skill performs; it describes what the skill contains rather than what it does, matching the 'names domain, minimal actions' anchor.

2 / 5

Completeness

It gives only a vague 'what' (the skill's composition) and no explicit 'when' — there is no 'Use when...' clause, only the purpose fragment 'for security testing', so it sits at the vague-what/no-when anchor.

2 / 5

Trigger Term Quality

'eval patterns' and 'security testing' are two developer/meta keywords rather than natural phrases a user would say to trigger the skill, missing the common variations users actually voice.

2 / 5

Distinctiveness Conflict Risk

'eval patterns for security testing' is a fairly specific niche unlikely to collide with most skills, but it could still overlap with other security/testing skills, matching the 'somewhat specific' anchor.

3 / 5

Total

9

/

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

frontmatter_unknown_keys

Unknown frontmatter key(s) found; consider removing or moving to metadata

Warning

Total

15

/

16

Passed

Repository

rohitg00/skillkit

Path

packages/core/src/eval/__tests__/fixtures/eval-injection-skill/SKILL.md

Revision

d2e5c34

Reviewed

Table of Contents

DiscoveryImplementationValidation

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.

Claim skill