Name: prompt-caching
Rating: 43.2 (1 reviews)
Author: administrakt0r

prompt-caching

You're a caching specialist who has reduced LLM costs by 90% through strategic caching. You've implemented systems that cache at multiple levels: prompt prefixes, full responses, and semantic similarity matches.

Quality

44%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

Securityby

Passed

No findings from the security scan

Fix and improve this skill with Tessl

tessl review fix ./plugins/AI-Agents-Safe-Coding-Skills-claude/skills/prompt-caching/SKILL.md

The canonical home for this skill is prompt-caching in administrakt0r/AI-Agents-Safe-Coding-Skills

Quality

Content

40%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 well-sectioned and brief but lacks any executable guidance: no code, no commands, a truncated principles list, and unexplained anti-patterns. It reads as an outline of a skill rather than a working skill file.

Suggestions

Add concrete, executable examples for each pattern (e.g. an Anthropic prompt-caching API call with cache_control markers) so the guidance is copy-paste ready.

Fix the truncated core-principles list ("2. K") and flesh out each anti-pattern with a one-line explanation of why it fails.

Replace the generic "When to Use" boilerplate with explicit trigger conditions and a short decision workflow for choosing prefix vs. response vs. semantic caching.

Dimension	Reasoning	Score
Conciseness	The body is short and mostly lean, but tokens like the repeated persona intro, the boilerplate "When to Use" line, and the broken principle "2. K" do not earn their place, placing it at the mostly-efficient-with-some-unnecessary-content anchor.	3 / 5
Actionability	There is no executable code or concrete command anywhere; guidance is limited to high-level hints ("Use Claude's native prompt caching for repeated prefixes") and bare capability tags, matching the minimal-concrete-guidance anchor.	2 / 5
Workflow Clarity	No sequenced workflow or validation checkpoints exist, and the single action is ambiguous rather than unambiguous, so the simple-skill exception does not apply; this sits at the rough-sequence-with-many-gaps anchor.	2 / 5
Progressive Disclosure	With no bundle files present the body is self-contained and organized into clear sections (Capabilities, Patterns, Anti-Patterns, Sharp Edges, Related Skills), but the empty anti-pattern subsections are a minor organization gap, matching the good-structure-with-minor-gaps anchor.	4 / 5
	Total	11 / 20 Passed

Description

48%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 establishes a clear niche (LLM cost reduction via multi-level caching) but relies on second-person persona framing and over-claims ("reduced LLM costs by 90%") instead of concrete actions, and it omits any explicit trigger guidance. It is distinguishable but not yet actionable as a trigger description.

Suggestions

Rewrite in third person naming concrete actions, e.g. "Caches LLM prompt prefixes, full responses, and semantic-similarity matches to reduce cost and latency."

Add an explicit "Use when…" clause with natural trigger terms such as "reduce LLM costs", "prompt caching", or "cache repeated prompts".

Drop the unverifiable persona claim ("reduced LLM costs by 90%") and second-person voice to lift specificity and avoid the voice penalty.

Dimension	Reasoning	Score
Specificity	Names the LLM-caching domain and lists three caching levels ("prompt prefixes, full responses, and semantic similarity matches"), but frames them as a persona ("caching specialist who has reduced LLM costs by 90%") rather than concrete actions; the second-person voice ("You're a…", "You've implemented…") applies the −1 specificity penalty, dropping it from a base of 3 to 2.	2 / 5
Completeness	Provides a (persona-flavored) "what" but no "Use when…" clause or equivalent trigger guidance; the rubric caps completeness at 3 when explicit trigger guidance is absent.	3 / 5
Trigger Term Quality	Contains relevant terms ("LLM costs", "caching", "prompt prefixes") but leans technical and misses the natural variations a user would actually say, matching the anchor for some relevant keywords missing common synonyms.	3 / 5
Distinctiveness Conflict Risk	LLM caching is a fairly distinct niche with only minor overlap risk against closely related skills like RAG or context-window management, matching the "mostly distinct" anchor.	4 / 5
	Total	12 / 20 Passed

Validation

93%

Warnings & errors only

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

Criteria	Description	Result
frontmatter_unknown_keys	Unknown frontmatter key(s) found; consider removing or moving to metadata	Warning

	Total	15 / 16 Passed

Repository: administrakt0r/AI-Agents-Safe-Coding-Skills
Path: plugins/AI-Agents-Safe-Coding-Skills-claude/skills/prompt-caching/SKILL.md
Commit: 57c135b

Reviewed: 2 days ago

Table of Contents

Discovery Implementation Validation

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.