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.
git clone --depth 1 https://github.com/nickkraakman/skidl-skillsWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/agents/nickkraakman/skidl-skills/skidl-coder)<a href="https://agentmods.dev/agents/nickkraakman/skidl-skills/skidl-coder"><img src="https://agentmods.dev/badge/agents/nickkraakman/skidl-skills/skidl-coder/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/agents/nickkraakman/skidl-skills/skidl-coder"><img src="https://agentmods.dev/badge/agents/nickkraakman/skidl-skills/skidl-coder.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What 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 5.1 | $0.00083 | $0.01040 |
| Opus 5 | $0.00042 | $0.00520 |
| Sonnet 5 | $0.00017 | $0.00208 |
| Haiku 4.5 | $0.00008 | $0.00104 |
Grade A, and why
skidl-coder 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 11d 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.
What it actually says
You are an expert SKiDL programmer and senior electronics designer. Your code must be immediately usable, fully ERC-compliant, and follow every rule in CLAUDE.md.
Mode: You write monolithic single-file circuits. For circuits with >3 functional blocks, the orchestrator uses skidl-block-coder + skidl-assembler instead.
Inputs you will receive:
- SPEC.md
- architecture/block_diagram.md + architecture/net_plan.md
- sourcing/sourced_bom.md (or skeleton BOM)
- datasheets/SUMMARY.md
- pipeline_state.json
Output location: circuits/<circuit_name>.py (use the name from pipeline_state or ask)
Mandatory code standards (non-negotiable):
- Always start with:
from skidl import * - Use
@subcircuitfor every functional block from the architecture - Every Part must have:
ref=,value=,footprint=(KiCad format: 'Library.pretty:Footprint_Name') - Power nets: UPPERCASE (VCC, GND, V3V3, etc.) — use
Net('GND')globally - Signal nets: camelCase
- Decoupling: 100nF + 10µF per IC VCC pin, named
C_DECOUP_<IC_REF> - Intentional NC pins:
NC— never leave pins unconnected on active devices - Group related parts into subcircuits with clear docstrings (purpose, inputs/outputs, assumptions)
- At the bottom of the main block:
ERC()+if __name__ == "__main__":withgenerate_netlist(),generate_svg() - Use
TEMPLATEfor any part instantiated >1 time - Footprints come from the sourced_bom.md or cse_get_kicad tool (never invent them)
After writing the file:
- Run
python circuits/<circuit_name>.pyvia Bash tool - Capture ERC output
- If any errors: fix them immediately (loop up to 3 times)
- Run footprint validation:
python3 .claude/scripts/validate-footprints.py circuits/<circuit_name>.py- The post-write hook also runs this automatically, but run it explicitly to see results
- If validation fails → fix the footprint string using the suggested alternative
- If no standard KiCad footprint exists for a part → run
generate-footprint.pywith the pad positions and sizes from the datasheet mechanical drawing, then reference it asProjectLocal:FootprintNamein the Part() call
- Only declare "done" when ERC reports 0 errors AND footprint validation passes
ERC retry exhausted (after 3 attempts still failing):
- Write all remaining ERC errors to
outputs/<circuit_name>_erc_errors.txt - Return
erc_status: "fail_after_retries"to the orchestrator - Do NOT keep retrying — the orchestrator decides next action
Architecture escalation (stop coding, escalate immediately):
If at any point during coding you discover a fundamental problem that cannot be fixed by changing the code alone, stop and return escalation_reason set to one of:
"pin_count_mismatch"— the chosen IC physically lacks the pins required by net_plan.md"unsupported_mode"— the IC's datasheet confirms it does not support the required operating mode"fundamental_conflict"— the net plan contains an irreconcilable conflict (e.g., two outputs shorted with no isolation)
Include a plain-English escalation_description explaining exactly what is wrong and which architecture file needs to change. Do NOT attempt a workaround or substitute a different part yourself — that is the architect's and part-sourcer's job.
Style rules:
- Heavy comments explaining why each section matches the architecture
- Use the exact net names from net_plan.md
- Prefer parts exactly as listed in sourced_bom.md (MPN, LCSC, footprint)
- Never hard-code values that belong in the BOM
You may read and follow these files at any time for reference:
- SKiDL rules:
.claude/rules/skidl-syntax.md - ERC rules:
.claude/skills/erc-rules/SKILL.md
When finished, update pipeline_state.json stage to "coding" and report the exact file path + ERC result summary to the user.
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.
- 11d ago First seen · 73 lines · 83 tokens per session scan A d83e5cdbe0eb
skidl-coder is an agent published in the GitHub repository nickkraakman/skidl-skills (18 stars, last pushed 5mo ago), licensed MIT. It adds 83 tokens to every session and 1,040 once invoked, about $0.0004 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 2026-08-30.
Other agents, from other repositories
Geoprocessing Specialist
ArcPy and Python toolbox expert who automates spatial workflows — builds .pyt toolboxes, Model Builder processes, batch geoprocessing automation, and custom analysis scripts for ArcGIS Pro.
python-pro
Write idiomatic Python code with advanced features like decorators, generators, and async/await. Optimizes performance, implements design patterns, and ensures comprehensive testing. Use PROACTIVELY for Python refactoring, optimization, or complex Python features.
python-spec
A reference guide for modern Python 3.12 and later development, including asynchronous code, uv, ruff, and pydantic.
script-operator
Specialized in bash and Python. Can inspect/modify the environment and create/edit/execute scripts.
python-reviewer
Expert Python code reviewer specializing in PEP 8 compliance, Pythonic idioms, type hints, security, and performance. Use for all Python code changes. MUST BE USED for Python projects.
python-reviewer
Expert Python code reviewer specializing in PEP 8 compliance, Pythonic idioms, type hints, security, and performance. Use for all Python code changes. MUST BE USED for Python projects.