Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx agentmods add skills/lhbsaa/embedded-dev-skill/embedded-implementationnpx skills add lhbsaa/embedded-dev-skill --skill embedded-implementationgit clone --depth 1 https://github.com/lhbsaa/embedded-dev-skillWhat it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5 | $0.00028 | $0.00921 |
| Opus 5 | $0.00014 | $0.00461 |
| Sonnet 5 | $0.00006 | $0.00184 |
| Haiku 4.5 | $0.00003 | $0.00092 |
Grade A, and why
embedded-implementation scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 2d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 163 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Embedded Implementation
Overview
Execute implementation plan with two-stage review after each task.
Core principle: Fresh subagent per task + hardware validation + code quality review = high quality.
Context: Run after embedded-driver-design creates plan.
Two-Stage Review Process
Task Execution → Hardware Validator → Code Quality Reviewer → Commit
↓ ↓ ↓
Generate code DMA/GPIO check MISRA C check
Stage 1: Hardware Validator
- Use
prompts/hardware-validator.md - Checks: DMA <= 4092, interface config, init sequence
Stage 2: Code Quality Reviewer
- Use
prompts/code-quality.md - Checks: MISRA C, DRY, YAGNI, clean code
Process
1. Load Plan
read docs/embedded/plans/YYYY-MM-DD-<driver>-plan.md
Extract ALL tasks with full text and context.
2. Create Todo List
todowrite todos=[{task: "Task 1: Header file", status: "pending"}, ...]
3. Execute Per Task
For each task:
#### Task N Execution
1. Mark todo: in_progress
2. Get task text and context (already extracted)
3. Dispatch implementation (use prompts/driver-generator.md)
- Generate code
- Build verification
4. Stage 1 Review: Hardware Validator
- Use prompts/hardware-validator.md
- Check DMA, interface, init sequence
5. If Stage 1 FAILS:
- Fix issues
- Re-review
- Don't proceed until APPROVED
6. Stage 2 Review: Code Quality
- Use prompts/code-quality.md
- Check MISRA C, DRY, YAGNI
7. If Stage 2 FAILS:
- Fix issues
- Re-review
- Don't proceed until APPROVED
8. Both stages APPROVED:
- Mark todo: completed
- Commit changes
4. After All Tasks
Dispatch final review for entire implementation
5. Transition to Verification
All tasks complete. Invoking embedded-verification for build-flash-monitor loop.
REQUIRED: Invoke embedded-verification next.
Implementation Prompt
Use prompts/driver-generator.md for code generation:
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 2d ago First seen · 163 lines · 28 tokens per session scan A 32e3a0ba8889
embedded-implementation is a skill published in the GitHub repository lhbsaa/embedded-dev-skill (9 stars, last pushed 1mo ago), licensed MIT. It adds 28 tokens to every session and 921 once invoked, about $0.0001 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other skills, from other repositories
agent-registry
Use this skill whenever the user wants to interact with Google Cloud's Agent Registry using gcloud commands. Triggers on any mention of "agent registry", "agent-registry", "mcp-servers", "gcloud agents", "register an agent", "list agents", "create a service", "agent service", or any request to manage agents, MCP…
html-reporting
Produce single-file, self-contained HTML artifacts instead of Markdown for any output a human will actually read. Use whenever the user asks for a "report", "research summary", "write-up", "spec", "plan", "PR explainer", "code review", "design mockup", "prototype", "dashboard", "deck", or anything that benefits from…
building-pydantic-ai-agents
Build AI agents with Pydantic AI — tools, capabilities, structured output, streaming, testing, and multi-agent patterns. Use when the user mentions Pydantic AI, imports pydanticai, or asks to build an AI agent, add tools/capabilities, stream output, define agents from YAML, or test agent behavior.
confluence-datacenter
This skill should be used when the user asks to "get Confluence page", "create Confluence page", "search Confluence", "update wiki page", or needs to interact with Confluence Data Center/Server instances.
github-pr-feedback
Use this skill whenever the user asks to look at GitHub PR feedback, PR review comments, review bot output, Gemini/Claude review feedback, or pasted reviewer suggestions and wants a markdown table that separates what should be fixed from what does not need to be fixed. This skill should trigger for PR feedback triage…
jira-datacenter
This skill should be used when the user asks to "get Jira issue", "create Jira ticket", "search Jira", "update issue status", or needs to interact with Jira Data Center/Server instances.