Name: context-compression
Rating: 44.8 (1 reviews)
Author: administrakt0r

context-compression

When agent sessions generate millions of tokens of conversation history, compression becomes mandatory. The naive approach is aggressive compression to minimize tokens per request.

Quality

46%

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/context-compression/SKILL.md

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

Quality

Content

57%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 content is information-rich and largely actionable with concrete templates and quantified comparisons, but it is verbose in places, lacks validation checkpoints in its workflows, and inlines material that would benefit from real one-level-deep reference files.

Suggestions

Add explicit validation checkpoints to the workflows, e.g. after each compression cycle run a probe-based recall/artifact check and only proceed if key details survive.

Trim explanation of concepts Claude already knows (basic compression trade-offs, ROUGE/embedding similarity) to keep the body lean.

Move the detailed evaluation framework and worked examples into actual reference files under references/ and link to them one level deep, since no bundle files currently exist.

Dimension	Reasoning	Score
Conciseness	The body is mostly efficient and carries domain-specific value (structured templates, comparison tables, specific thresholds), but it also explains concepts Claude already knows — e.g. "Context compression trades token savings against information loss", what ROUGE/embedding similarity are, and basic sliding-window intuition — which could be trimmed.	3 / 5
Actionability	It provides a copy-paste-ready structured summary template, concrete trigger thresholds (70-80%), comparison tables with specific ratios and quality scores, and a numbered implementation procedure for anchored iterative summarization; only minor specifics (e.g. how to merge sections) are left implicit.	4 / 5
Workflow Clarity	The three-phase workflow and the five-step implementation sequence are clearly ordered, but neither embeds explicit validation/verification checkpoints (validate-then-proceed, feedback loops), which the rubric requires for destructive/information-loss operations like compression.	3 / 5
Progressive Disclosure	The body is well sectioned, but it is a 258-line monolith far above the simple-skill threshold; bulk content that belongs in separate files (the full evaluation framework, the worked examples) is inlined, and the only "Internal reference" named points to no real file (no references/ directory exists).	3 / 5
	Total	13 / 20 Passed

Description

36%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 reads as a thesis statement about a problem rather than a crisp capability+trigger declaration. It lacks concrete actions, a "Use when…" clause, and natural trigger phrases, leaving it below the rubric's good examples.

Suggestions

Rewrite the description to lead with concrete capabilities, e.g. "Summarizes long agent conversation history into structured persistent summaries tracking files modified, decisions, and next steps."

Add an explicit "Use when…" trigger clause listing natural phrases such as "session exceeds context window", "agent forgets modified files", or "compressing conversation history".

Tighten distinctiveness by naming the specific technique (anchored iterative summarization) and the artifact-trail focus that separates this skill from general context-optimization.

Dimension	Reasoning	Score
Specificity	The description names the domain ("compression becomes mandatory") and gestures at an action ("aggressive compression to minimize tokens per request") but never lists concrete capabilities of the skill; it frames a problem rather than stating what the skill does.	2 / 5
Completeness	It offers only a vague "what" (the problem of compression) and contains no "Use when…" or equivalent trigger clause; the missing explicit trigger guidance would cap completeness at 3, but the "what" itself is too vague to reach that.	2 / 5
Trigger Term Quality	Relevant keywords like "agent sessions", "conversation history", "compression", and "tokens" appear, but common variations and synonyms a user would naturally say (e.g. "context window", "summarize session", "context too long") are missing.	3 / 5
Distinctiveness Conflict Risk	It is tied to context compression for agent sessions, which is fairly specific, but the framing overlaps noticeably with related context-engineering skills (context-optimization, context-degradation, memory-systems) it itself references.	3 / 5
	Total	10 / 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/context-compression/SKILL.md
Commit: 57c135b

Reviewed: 3 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.