bedrock-inference

Amazon Bedrock Runtime API for model inference including Claude, Nova, Titan, and third-party models. Covers invoke-model, converse API, streaming responses, token counting, async invocation, and guardrails. Use when invoking foundation models, building conversational AI, streaming model responses, optimizing token usage, or implementing runtime guardrails.

Quality

73%

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 ./skills/bedrock-inference/SKILL.md

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.

An actionable, code-rich reference for Bedrock inference that is let down by verbosity and a monolithic structure. It advertises progressive disclosure to reference files that are not bundled, and its batch paths lack validation.

Suggestions

Actually create the referenced bundle files (references/advanced-invocation.md, multimodal.md, tool-use.md, performance.md, monitoring.md) and move the corresponding detail out of SKILL.md, or remove the dead references.

Trim non-load-bearing sections — 'Quality Targets', the dated pricing table, and inference-profile prefix explanations — since Claude can look up current pricing and profile semantics at runtime.

Add a validation/feedback checkpoint to the async batch flow (verify each invocation ARN succeeded, handle Failed status, retry) so batch operations clear the workflow-clarity cap.

Dimension	Reasoning	Score
Conciseness	The ~1000-line body is mostly code but carries notable non-essential padding — 'Quality Targets' (latency/throughput/SLA), a hardcoded pricing table 'as of December 2024', and inference-profile prefix explanations Claude could look up — placing it at 'mostly efficient but includes some unnecessary explanation' rather than the lean score 4.	3 / 5
Actionability	Code is largely copy-paste ready and covers the common cases (invoke, converse, stream, count, async, guardrails, IAM), but key flows leave placeholder completions such as '(Add tool result to messages and call converse again)' and '(Use boto3 S3 client to retrieve)', the minor gaps that hold it below 5.	4 / 5
Workflow Clarity	Sequences exist for async (start → check status → list) and error handling has retry/backoff feedback loops, but batch usage ('use_async_for_batch') iterates prompts with no validation of results, so the batch-operation cap of 3 applies.	3 / 5
Progressive Disclosure	A dedicated 'Progressive Disclosure' section signals one-level-deep references, but the referenced files (references/advanced-invocation.md, multimodal.md, tool-use.md, performance.md, monitoring.md) do not exist, and the bulk reference content is inlined in the monolithic SKILL.md rather than split out.	3 / 5
	Total	13 / 20 Passed

Description

88%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.

A strong, third-person description that clearly states the capability surface and gives explicit 'Use when' triggers. It is comprehensive and actionable, with only minor overlap risk against closely related Bedrock skills.

Dimension	Reasoning	Score
Specificity	Lists six concrete actions — 'invoke-model, converse API, streaming responses, token counting, async invocation, and guardrails' — giving comprehensive coverage of the Bedrock Runtime API surface, matching the score-5 anchor for multiple specific concrete actions.	5 / 5
Completeness	Explicitly answers both what ('Covers invoke-model, converse API, streaming responses...') and when ('Use when invoking foundation models, building conversational AI...') with concrete trigger phrases, matching the score-5 anchor.	5 / 5
Trigger Term Quality	Natural phrases like 'streaming model responses', 'building conversational AI', and 'optimizing token usage' give good coverage, but the triggers lean technical ('runtime guardrails', 'invoking foundation models') and lack common synonyms, fitting score 4 rather than the comprehensive 5.	4 / 5
Distinctiveness Conflict Risk	The Bedrock Runtime inference niche is clearly distinct, but the description touches guardrails and token/cost optimization which overlap with sibling skills (bedrock-guardrails, claude-cost-optimization), so it is mostly distinct with minor overlap rather than minimal conflict risk.	4 / 5
	Total	18 / 20 Passed

Validation

75%

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 — 12 / 16 Passed

Validation for skill structure

Criteria	Description	Result
skill_md_line_count	SKILL.md is long (1128 lines); consider splitting into references/ and linking	Warning
allowed_tools_field	'allowed-tools' contains unusual tool name(s)	Warning
relative_links	Relative link issues: 5 missing	Warning
referenced_paths_exist	Referenced path issues: 5 missing	Warning

	Total	12 / 16 Passed

Repository: fernandezbaptiste/Skrillz
Path: skills/bedrock-inference/SKILL.md
Commit: 93ed392

Reviewed: 1 day 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.