Skip to content

feat: add bedrock plugin for Amazon Bedrock setup and operations#127

Open
balak4 wants to merge 1 commit intoawslabs:mainfrom
balak4:feature/bedrock-plugin
Open

feat: add bedrock plugin for Amazon Bedrock setup and operations#127
balak4 wants to merge 1 commit intoawslabs:mainfrom
balak4:feature/bedrock-plugin

Conversation

@balak4
Copy link
Copy Markdown

@balak4 balak4 commented Apr 10, 2026

Authors

Bala Krishnamoorthy, Sr. GenAI Data Scientist, Bedrock GTM

Summary

  • Add bedrock plugin v1 providing guided Amazon Bedrock setup — IAM configuration, model access, prompt caching, observability, quota optimization, and cost analysis
  • 8 slash commands, 6 validation scripts, 1 SKILL.md with 8 reference docs, and AWS Documentation MCP server integration
  • Introduces commands/ directory pattern for intent-based routing via /bedrock entry point

What's Included

Component Count Description
Commands 8 /bedrock router + 7 sub-commands (setup, cache, cache-debug, quota, usage, costs, validate-model-access)
Scripts 6 Shell and Python validation/analysis scripts (boto3 only, no extra dependencies)
Skill 1 SKILL.md with auto-trigger on Bedrock-related intent
References 8 IAM permissions, profile setup, model access, prompt caching, observability, quota optimization, cost optimization, quick reference
MCP Server 1 aws-documentation for live AWS docs lookup

How It Works

The /bedrock command acts as a natural-language router — it matches developer intent to the appropriate sub-command or knowledge source. All AWS operations require explicit --profile selection (instruction-based enforcement, no hooks). Scripts make real AWS API calls for validation (~$0.01–0.08 per run).

Testing

All 6 scripts validated against a live AWS account with default profile:

  • validate-bedrock-access.sh — 4/4 checks pass
  • validate-prompt-caching.py — cache write + read confirmed
  • check-quota-health.py — quota analysis completes with recommendations
  • debug-prompt-cache.py — 6/6 diagnostic tests pass
  • analyze-bedrock-usage.py — usage report generates
  • analyze-bedrock-costs.py — cost report generates

Repo validations:

  • All JSON schemas pass (marketplace, plugin, MCP, codex-plugin, codex-marketplace)
  • Cross-references: 0 errors, 0 warnings
  • SKILL.md size: "ideal" (185 lines, 3451 tokens)
  • Markdown lint: 0 new errors
  • Bandit security scan: 0 issues
  • dprint formatting: clean

By submitting this pull request, I confirm that you can use, modify, copy, and redistribute this contribution, under the terms of the project license.

@theagenticguy
Copy link
Copy Markdown
Contributor

Hi @balak4, thank you for the thorough work here — the level of detail in the implementation and testing is really impressive!

After reviewing this, we think the Bedrock setup and operations functionality would be a better fit as part of a broader AWS plugin rather than a standalone plugin. There's an active RFC in #101 (Work With AWS) that's shaping up to be the umbrella for this kind of AWS-specific tooling, and Bedrock support would be a great addition to that effort.

We'd love for you to share your ideas and this work over on that RFC discussion — your expertise here would be really valuable in shaping the direction of that plugin. Per our Contributing Guidelines, new plugins should go through the RFC process first, which helps ensure alignment before implementation work begins.

Thanks again for your interest in contributing, and we hope to see you involved in the RFC discussion!

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants