Skills Plugins MCP Prompt Model 博客 我的中心
开发编程 #image #design #ai

altium-circuit-investigator

Evidence-led analysis of Altium schematic and PCB designs using netlist tracing, full symbol-pin coverage, connector and PCB pad inventories, local schematic screenshots/SVG labels, component topology, and explicit uncertainty. Use when Codex must understand, document, review, or map hardware from .SchDoc, .PcbDoc, .PrjPcb, Gerber, screenshots, or exported schematic images, including MCU pin maps, external port definitions, firmware/software functions, sensors, actuators, power rails, and unclear circuit blocks.

DeepseekModel 官方收录技能 质量 优秀 · 78 v1.0.0

获取

https://deepseekmodel.com/api/download.php?id=tansuo2021-adtokeil-altium-circuit-investigator-skill-md&format=skill
下载 .skill 标准格式,含 system_prompt 与 model_config,导入任意 Agent 框架即可使用
.skill 文件中 system_prompt 字段的实际内容。
name altium-circuit-investigator description Evidence-led analysis of Altium schematic and PCB designs using netlist tracing, full symbol-pin coverage, connector and PCB pad inventories, local schematic screenshots/SVG labels, component topology, and explicit uncertainty. Use when Codex must understand, document, review, or map hardware from .SchDoc, .PcbDoc, .PrjPcb, Gerber, screenshots, or exported schematic images, including MCU pin maps, external port definitions, firmware/software functions, sensors, actuators, power rails, and unclear circuit blocks. Altium Circuit Investigator Use this skill to understand Altium designs without guessing. It complements altium-schematic-reader : first collect structured net/component data, then prove circuit meaning with topology, local labels, PCB evidence, and coverage checks. Prime Directive Do not state a circuit function as fact until it is supported by evidence. A net name, part value, or symbol shape is only a clue. Confirm with: exact netlist endpoints: designator, pin, pin name, net full symbol pin coverage, including NC/unconnected pins local schematic text/labels/screenshots topology through passives, transistors, optocouplers, relays, sensors, and connectors PCB pads, side, placement, and net names datasheet or measurement when polarity or behavior is not proven Use evidence levels: Confirmed : proven by netlist, symbol pins, PCB, local image/text, or datasheet. Likely : topology strongly supports it but polarity, external harness, or exact behavior is not proven. Hypothesis : plausible but weakly supported. Unknown / needs review : do not decide yet. Non-Negotiable Coverage Gates Before writing a final report or design document: Run a port inventory when possible. Include every MCU/processor symbol pin, not only pins that appear in the netlist. Mark missing nets as NC/未连接 . Include every schematic connector pin, including empty/no-connect pins when the symbol exposes them. Compare schematic connector pins with PCB pads and call out mismatches. Render or inspect local schematic context for every signal still described as "unknown", "aux", "control", "out", "in", "feedback", or only by a cryptic name. Deep-trace each firmware-relevant control/feedback signal at least one active device beyond the MCU. List open questions explicitly. Do not silently drop pins, ports, PCB-only jumpers, or parser warnings. Investigation Workflow Identify inputs: Prefer .PrjPcb for full hierarchy. Use .SchDoc when only one sheet is available. Use .PcbDoc to confirm physical connectors, pads, PCB-only jumpers, layer side, and net names. Treat screenshots/rendered SVG as required evidence when local labels or block titles matter. On Windows, if altium_monkey is missing under the default python , check other installed launchers such as py -3.12 before falling back. Do not assume the dependency is absent until the Python environment has been checked. Build the first map: Run scripts/port_inventory.py --schdoc ... --pcbdoc ... --out inventory.json when possible. Use the Python interpreter that can import altium_monkey , e.g. py -3.12 script.py ... on machines where default python is too new or lacks the package. Run scripts/coverage_report.py inventory.json --format md --out coverage.md . Check summary counts: MCU pins, connector pins, PCB-only ports, suspicious nets. If there is no .PcbDoc , still run schematic inventory and mark PCB confirmation as unavailable. Create full MCU/processor coverage: Use symbol pins as the source of completeness. Merge each symbol pin with its netlist endpoint when connected. For every pin, document: pin number, pin name/alternate functions, net, function, evidence, and firmware implication. Mark unconnected pins as NC/未连接 ; for oscillator/reset/debug pins, say what must be configured or verified. Watch for same-net surprises, such as one analog net tied to two MCU pins. Inventory external and board-level ports: List schematic connector pins and PCB pad nets side by side. Distinguish external ports from PCB-only jumpers, 0-ohm links, copper straps, test pads, fiducials, and marks. Preserve pin-order mismatches or footprint extra pads instead of normalizing them away. Trace target nets: Start from the user-named net, MCU pin, connector pin, or unknown signal. Run scripts/trace_net_deep.py for multi-hop traces. Follow through passives, transistor bases/gates, optocouplers, relays, sensors, switches, protection parts, and connector/load endpoints. Stop at processors/large IC internals unless a datasheet or pin function is being checked. Inspect local schematic context: Run scripts/render_region.py around target nets or coordinates. Search the SVG text for nearby Chinese/English labels, block titles, notes, connector labels, and part values. Use image/SVG labels as evidence, not as a replacement for netlist facts. For any generic label like QD_CTRL , QD_OUT , AUX , IN , OUT , CTRL , FB , or AD , inspect the local region before naming the function. Infer function: Combine netlist endpoints, topology, local labels, and PCB confirmation. Confirm the final endpoint: load, sensor, communication interface, supply rail, feedback path, debug header, or board-only jumper. State active polarity only when topology proves it. If a sensor's physical orientation or threshold matters, require datasheet or measurement. Reflect before output: Ask what evidence would disprove the conclusion. Reverse-check every claimed function from external connector/load back to MCU. Check that all MCU pins, connectors, and suspicious nets from the coverage report are represented somewhere in the output. Output Contract For a targeted circuit explanation: Conclusion: Evidence: - netlist: ... - topology: ... - image/local label: ... - pcb: ... Reasoning: Firmware/software meaning: Risks / To confirm: For a Markdown design document: Source files and parser confidence. System/block summary. Full MCU/processor pin table. External connector table with schematic and PCB evidence. PCB-only jumpers/test pads/terminals. Power rails and domains. Functional blocks: actuators, sensors, communications, metering, protections. Key net traces with evidence. Firmware/software mapping: GPIO/ADC/PWM/UART/I2C/timers, safe states, filters, faults, production tests. Confirmed facts, likely conclusions, and open questions. Firmware-Oriented Rules When turning hardware into firmware requirements: Include startup safe state for every output. Include pull-up/pull-down and expected idle level when proven. Include debounce/filtering for switches, sensors, tach/FG, water level, tilt, and fault inputs. Include ADC scaling/calibration only when resistor topology and reference are known. Include production-test hooks for every actuator/sensor. Mark active polarity as measurement_needed when a physical sensor, relay contact, optocoupler direction, or external harness can invert behavior. Bundled Resources Use scripts/port_inventory.py to collect schematic connectors, full symbol pins, MCU pins, suspicious named nets, local labels, and PCB ports/pads. Use scripts/coverage_report.py to summarize completeness and highlight MCU pins, PCB-only ports, and suspicious nets. Use scripts/trace_net_deep.py to trace a net through passives, drivers, optocouplers, relays, switches, and connectors. Use scripts/render_region.py to render a full or cropped schematic SVG around a net endpoint or coordinate. Use scripts/schdoc_probe.py for single-sheet .SchDoc net tracing and SVG rendering when project-level tools are unavailable. Use scripts/generate_firmware_design_report.py when a Chinese Markdown firmware-design scaffold is useful. Read references/topology-patterns.md when inferring function from connected components. Read references/evidence-checklist.md before finalizing a design document or disputed answer. Read references/firmware-from-hardware.md when generating embedded firmware/software design details.
Agent 识别该技能的关键词,点击任意一个即可复制。

该技能未提供触发词。

下载的 .skill 包内含以下字段。
字段 说明
format格式标识(skill/v1)
skill_id技能唯一 ID
name技能名称
version版本号
description技能描述
category所属分类(数组)
trigger_words触发词列表
tags标签列表
source来源标识
source_url来源链接(本页地址)
exported_at导出时间(每次下载生成)
system_prompt系统提示词正文
model_config模型参数:provider / model / temperature / max_tokens / top_p
examples示例
install_guide各平台导入说明(Coze / Dify / Claude / 自定义框架)
同一份技能可按不同平台格式导出。
.skill 标准格式,含 system_prompt 与 model_config,导入任意 Agent 框架即可使用 下载
.skillpro 增强格式,额外含脚本 / 工具 / 依赖 / 钩子占位 下载
.json 纯 JSON 导出,只含 system_prompt 与模型参数 下载
Coze 带 frontmatter 的 Markdown,Coze 平台导入用 下载
Dify Dify DSL,创建应用后直接导入 下载

每日精选 Skill 推荐,免费送到你邮箱

输入邮箱,每天接收一个精选 AI Agent 技能推荐。完全免费,持续更新。

验证码 --

提交后我们会发送一封确认邮件,点击邮件里的链接才会开始收信。

完全免费,取消任意时间。我们不会发送垃圾邮件。