CtrlK
BlogDocsLog inGet started
Tessl Logo

iterative-retrieval

逐步优化上下文检索以解决子代理上下文问题的模式

49

Quality

54%

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 ./docs/zh-CN/skills/iterative-retrieval/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.

The body is a well-structured, actionable guide with executable code across all four phases, a clear looping workflow with a stop condition, and sensible one-level references. Its main gap is the absence of explicit validation/error-recovery checkpoints in the loop and a few undefined helper functions in the examples.

Suggestions

Add an explicit validation checkpoint in the loop (e.g., verify retrieved files actually exist and are parseable before scoring, and surface a clear 'retrieval failed' path) to push workflow clarity toward the explicit-validation anchor.

Either define or note the assumed helper functions (`retrieveFiles`, `scoreRelevance`, `explainRelevance`, `identifyGaps`, `extractPatterns`, `mergeContext`) so the code is fully copy-paste executable rather than relying on implied implementations.

Trim the ASCII diagram and the high/medium/low/none scoring-band list, or move them to a reference file, to tighten conciseness further.

DimensionReasoningScore

Conciseness

The body is efficient with code that earns its place and minimal padding of known concepts; only minor prose (the diagram and scoring-band list) could be trimmed, sitting clearly above the 'mostly efficient' anchor but short of fully lean.

4 / 5

Actionability

Provides concrete, executable JavaScript for each of the four phases plus a complete `iterativeRetrieve` function and two worked examples, with only minor gaps (undefined helper functions like `retrieveFiles`, `scoreRelevance`) keeping it from copy-paste completeness.

4 / 5

Workflow Clarity

The 4-phase dispatch→evaluate→refine→loop sequence is clearly staged with an explicit stop condition (≥3 high-relevance files) and a built-in feedback loop; it lacks explicit validation/error-recovery checkpoints, so it is above 'steps listed, validation gaps' but below the explicit-validation anchor.

4 / 5

Progressive Disclosure

Content is well-sectioned (when, problem, solution, examples, integration, best practices, related) with one-level references (external guide link, sibling `continuous-learning` skill, `agents/` path); no bundle files exist to over-fragment, so structure is good with only minor organization gaps.

4 / 5

Total

16

/

20

Passed

Description

33%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 conveys a clear 'what' but omits any explicit 'when to use' trigger guidance and uses technical rather than natural user phrasing. It is a single-domain pattern description with moderate distinctiveness but weak trigger-term quality and completeness.

Suggestions

Add an explicit 'Use when...' clause listing natural trigger phrases (e.g., 'Use when a sub-agent lacks needed codebase context, or when retrieval hits context-too-large / missing-context failures').

Include natural user-side keywords and synonyms alongside the technical terms (e.g., 'context retrieval', 'RAG for agents', 'sub-agent context', 'code exploration') so the description matches how users actually ask.

Spell out the concrete actions the pattern provides (e.g., 'dispatch broad queries, score file relevance, refine search terms, loop up to 3 cycles') to lift specificity from a single abstract action to multiple concrete ones.

DimensionReasoningScore

Specificity

Names the domain ("上下文检索" / context retrieval for sub-agents) and a concrete action ("逐步优化" / iteratively refining), but as a single "pattern" description it lists only one broad action rather than multiple specific ones, matching the 'domain and 1-2 concrete actions' anchor.

3 / 5

Completeness

It states a clear 'what' (patterns for iteratively refining context retrieval) but entirely lacks a 'when'/'Use when...' trigger clause; per the missing-trigger cap this cannot exceed 3, and with no implied 'when' it sits at 2.

2 / 5

Trigger Term Quality

Terms are technical ("上下文检索", "子代理上下文问题") rather than natural phrases a user would say; no common synonyms or user-side trigger phrasing, sitting between generic and 'some relevant keywords'.

2 / 5

Distinctiveness Conflict Risk

The sub-agent context-retrieval niche is somewhat specific but could still overlap with general RAG/retrieval skills, matching the 'somewhat specific but could still overlap' anchor.

3 / 5

Total

10

/

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.

Validation15 / 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
affaan-m/ECC
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.