CtrlK
BlogDocsLog inGet started
Tessl Logo
SkillPublic

cache-deception

Web cache deception — trick CDN/proxy into caching authenticated responses under unauthenticated URLs, exposing PII to any visitor.

Quality

73%

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/decepticon/decepticon/skills/standard/exploit/web/cache-deception/SKILL.md
SKILL.md
Quality
Evals
Security

Quality

Content

75%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.

A well-structured, token-efficient offensive-security skill with concrete probe commands, a variant matrix, a runnable PoC with explicit confirmation criteria, and defender guidance. The main gaps are an undefined helper command/placeholders in the probe steps and the absence of an explicit error-recovery loop when a suffix probe fails.

Suggestions

Replace the undefined 'diff_response_bodies' with a concrete command (e.g., a curl | diff one-liner) and show how $TARGET/$AUTH are set so both probe steps are copy-paste ready.

Add a feedback loop after step 2: if the unauthenticated replay misses or returns 404, iterate the §3 variant table and retry, converting the variants list into an explicit error-recovery path.

Move the dated 'Known exemplars' content into a separate reference file or a clearly framed context section to keep the main body evergreen.

DimensionReasoningScore

Conciseness

The body is lean and table/code-driven and assumes Claude's competence (no primer on what a CDN is), but has minor trimmable spots: 'This skill covers deception. Cache poisoning is a separate (overlapping) class' restates the bullets above it, and the 'Known exemplars' section carries dated bounty figures without an 'old patterns' framing.

4 / 5

Actionability

Mostly executable guidance: copy-paste curl probes with headers to inspect, a runnable Python PoC with asserts, a variant path table, and concrete nginx/Cloudflare defender config. Minor gaps: 'diff_response_bodies' is an undefined helper command, $TARGET/$AUTH are never defined, and the referenced probe site is unverifiable.

4 / 5

Workflow Clarity

Clear two-step probe sequence with explicit validation checkpoints (X-Cache MISS→HIT, Content-Length consistency, unauth replay confirmation, PoC asserts). Falls short of a 5 because there is no error-recovery loop — what to try when the second request misses or returns 404 (e.g., iterating the §3 variants) is only implicit.

4 / 5

Progressive Disclosure

No bundle files exist; the ~110-line body is well-sectioned with a clearly signaled 'Cross-references' section pointing one level deep (though to corpus paths outside this bundle). Severity ratings and dated exemplar content are inlined rather than split into a reference file, leaving minor organization gaps.

4 / 5

Total

16

/

20

Passed

Description

61%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 states a specific, correctly-scoped capability with a clear mechanism and impact, but omits any 'when to use' trigger guidance, which caps completeness and limits its effectiveness for skill selection. Trigger term coverage is good though not comprehensive.

Suggestions

Append an explicit trigger clause, e.g. 'Use when testing web apps behind a CDN or reverse proxy (Cloudflare, Fastly, Akamai, Varnish) for cache misconfigurations, or when the user mentions cache deception/poisoning.'

Enumerate the concrete actions the skill performs (probe URL suffix variants, replay unauthenticated to confirm, rate severity) so specificity reaches several listed actions rather than one core technique.

Include common synonyms users say ('cache deception attack', 'CDN cache misconfiguration') alongside the current keywords to widen natural trigger coverage.

DimensionReasoningScore

Specificity

The description names the domain and one concrete action with mechanism and outcome ("trick CDN/proxy into caching authenticated responses under unauthenticated URLs, exposing PII"), but covers only the single core technique rather than several listed actions, matching the '1-2 concrete actions, not comprehensive' anchor.

3 / 5

Completeness

The 'what' is clear and specific, but the description contains no 'Use when...' clause or equivalent trigger guidance, which caps completeness at 3 per the judging guidelines. Not a 2 because the 'what' is concrete, not vague.

3 / 5

Trigger Term Quality

Good natural keyword coverage — 'web cache deception', 'CDN', 'proxy', 'caching', 'PII' — terms a user would plausibly say; a few natural variations (cache poisoning, Cloudflare/Akamai, cache misconfiguration) are missing from the description itself (they appear only in metadata.when_to_use).

4 / 5

Distinctiveness Conflict Risk

'Web cache deception' is a niche with distinct vocabulary and minimal overlap risk with other skills, though the related-but-separate cache-poisoning trigger space is not disambiguated within the description. Anchor 5 exemplars include explicit 'Use when...' triggers, which this description lacks.

4 / 5

Total

14

/

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

PurpleAILAB/Decepticon

Path

packages/decepticon/decepticon/skills/standard/exploit/web/cache-deception/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.