[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M7FCM2CESJH9P38MBE14PHA0":3,"skill-same:01M7FCM2CESJH9P38MBE14PHA0":75},{"skillNo":4,"name":5,"nameZh":6,"description":7,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":16,"categories":17,"keywords":18,"publishedAt":19,"linkOnly":20,"markdownContent":21,"markdownBody":22,"frontMatter":23,"refName":27,"commitSha":28,"skillDirectory":29,"forksCount":30,"repositoryHtmlUrl":31,"repositoryHomepageUrl":6,"ownerHtmlUrl":32,"ownerBlogUrl":6,"ownerDescription":33,"updatedAt":34,"install":35,"resources":48},"01M7FCM2CESJH9P38MBE14PHA0","telegram-e2e-userbot",null,"Prove user-visible OpenClaw Telegram behavior on Telegram's Test Server with Convex-leased team credentials; drive real-user turns and record messages, edits, deletions, reactions, typing, or rich content.","lza6\u002FArmor-piercing-system-prompts","lza6","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F128880206?v=4","01M7EB3K7WPX5T28EJ2NVHE6EG","Chinese-tingfeng","01M7EB3K97E1KQ16QWKJMFNGX8",11,"MIT","https:\u002F\u002Fgithub.com\u002Flza6\u002FArmor-piercing-system-prompts\u002Fblob\u002Fmain\u002F06-前端与设计\u002Ftelegram-e2e-userbot\u002FSKILL.md",[],[],"2026-11-16 10:05:34",false,"---\nname: telegram-e2e-userbot\ndescription: \"Prove user-visible OpenClaw Telegram behavior on Telegram's Test Server with Convex-leased team credentials; drive real-user turns and record messages, edits, deletions, reactions, typing, or rich content.\"\nmetadata:\n  short-description: Telegram E2E via real-user driver\n  argument-hint: \"\u003Cmessage-or-command?>\"\n---\n\n# Telegram E2E (Userbot)\n\nProve OpenClaw behavior against Telegram's Test Server as a dedicated QA user.\nThe user driver sees edits, deletions, reactions, and typing that a second bot\ncannot observe.\n\nTDLib is the headless user transport. Evidence is a structured Telegram event\ntimeline.\n\nEach credential contains one SUT bot and one independent TDLib authorization\nfor the shared QA user identity. Independent authorizations keep parallel runs\nfrom splitting Telegram updates between observers.\n\nPool creation, session repair, and credential publication remain owner-only\noperations outside this repository skill.\n\nConvex controls allocation only. During a lease, the worker may change the\ntest config, edit this harness, run arbitrary argv commands, and use the leased\nTDLib session or any Telegram Test Bot API method.\n\n## 1. Prepare\n\nRun from the OpenClaw checkout and ref under test. Point to the repository skill:\n\n```bash\nTELEGRAM_E2E_SKILL_DIR=\"${TELEGRAM_E2E_SKILL_DIR:-$PWD\u002F.agents\u002Fskills\u002Ftelegram-e2e-userbot}\"\nexport TELEGRAM_E2E_SKILL_DIR\n```\n\nThe runner starts the built `dist\u002Fentry.js` from this checkout. Before acquiring a\ncredential, verify that the exact ref has a dependency-ready runtime and usable\nbuild output without an implicit install or build. A bare worktree with missing\ndependencies or a missing or stale build stamp is not ready. Move to a trusted\ndependency-ready checkout or exact-head artifact first.\n\nOn hosts where builds are prohibited, `--source-gateway` runs a dependency-ready\nTypeScript checkout through the repository's existing development launcher.\n\nUse two unused ports for every run on a shared host:\n\n```bash\n: \"${TELEGRAM_GATEWAY_PORT:?set an unused Gateway port}\"\n: \"${TELEGRAM_MOCK_PORT:?set an unused provider port}\"\nexport TELEGRAM_GATEWAY_PORT TELEGRAM_MOCK_PORT\n```\n\nTeam and ClawSweeper runs use Convex. A maintainer with Convex CLI access to the\nOpenClaw broker project needs no local broker settings. The lease helper uses\n`qa\u002Fconvex-credential-broker` in the checkout, reads the production site and\nCI role through the authenticated CLI, and keeps them in process memory. Run\nthe same doctor and runner commands below; no credential export is required.\nRequire an authenticated `convex` command on `PATH`. When it is missing or\ncannot access the broker project, stop and ask the user to install and\nauthenticate the Convex CLI. Runtime setup must never install or log in.\n\nCI or another non-interactive worker can instead provide the broker pair:\n\n```bash\n: \"${OPENCLAW_QA_CONVEX_SITE_URL:?missing Convex QA site}\"\n: \"${OPENCLAW_QA_CONVEX_SECRET_CI:?missing Convex QA CI credential}\"\n```\n\nDone when the checkout contains `scripts\u002Fe2e\u002Fmock-openai-server.mjs`, the skill\ndirectory resolves, the exact runtime can start without setup work, and either\nthe Convex CLI can access the production broker project or the broker pair is\npresent.\n\n## 2. Select the proof\n\nRead [the verification map](features\u002FREADME.md), then read only the recipe for\nthe behavior under test. Prefer a DM for isolation. Use the shared group only\nwhen group policy, mentions, commands, topics, or reactions are part of the\nclaim.\n\nExercise the exact behavior changed by the diff. The generic `OPENCLAW_E2E_OK`\nturn proves only the default message path. Formatting, commands, media, edits,\ndeletions, reactions, topics, and timing each need actions and recorded events\nthat expose the specific claim.\n\nTreat the harness as an extension point. When its current actions, recorder\nfields, or recipes cannot expose the claim, extend them in the tested checkout.\nUse the leased TDLib session or any Telegram Test Bot API method needed to cover\nthe behavior.\n\nFor a non-default backend or timed scenario, read the matching section of the\n[runtime reference](features\u002Fruntime-reference.md).\n\nUse the scenario `command` action when the proof needs custom setup, inspection,\nor a Telegram API call. It receives the leased test bot, TDLib session, local\nTest Bot API proxy, Gateway config, and Gateway state.\n\nDone when the prompt, chat type, config patch, and expected event sequence each\nmap to the behavior being proved.\n\n## 3. Run the doctor\n\n```bash\nnode \"$TELEGRAM_E2E_SKILL_DIR\u002Fscripts\u002Ftelegram-test-doctor.mjs\"\n```\n\nRequire `ok: true`. The doctor leases one credential, restores its private TDLib\nstate, confirms the pinned TDLib version on the Test Server, verifies the SUT\nthrough the Test Bot API proxy, then removes the state and releases the lease.\n\nDone when the doctor reports Convex source, loaded credentials, isolated TDLib\nstate, Test Server, authorized user, Bot API proxy, disabled group privacy,\nactive group membership, and the SUT bot.\n\n## 4. Drive and record\n\nCreate a durable proof directory outside runner scratch state:\n\n```bash\nTELEGRAM_E2E_PROOF_DIR=\"$(mktemp -d \u002Ftmp\u002Ftelegram-e2e-proof.XXXXXX)\"\nnode \"$TELEGRAM_E2E_SKILL_DIR\u002Fscripts\u002Frun-mock-sut-user-e2e.mjs\" \\\n  --gateway-port \"$TELEGRAM_GATEWAY_PORT\" --mock-port \"$TELEGRAM_MOCK_PORT\" \\\n  --dm --text 'Please answer with OPENCLAW_E2E_OK only.' \\\n  --record \"$TELEGRAM_E2E_PROOF_DIR\u002Fevents.ndjson\" \\\n  --output \"$TELEGRAM_E2E_PROOF_DIR\u002Fsummary.json\"\n```\n\nThe runner owns one Convex lease, the Test Bot API proxy, a\nfresh gateway, the selected provider, the real-user action, recording, teardown,\nand cleanup. The default gateway and mock-provider ports are `19879` and `19882`;\nmanaged workloads on one host always pass distinct `--gateway-port` and\n`--mock-port` values instead of using defaults.\n\nConvex selects and exclusively leases any free credential. An exhausted pool\nwaits briefly for a release. The runner heartbeats during the run and releases\nthe credential during cleanup.\n\nRecording captures facts and rejects probe assertions such as `--expect` and\n`--any-sut-reply`. Use probe mode only for a quick reply check.\n\nDone when the runner exits zero and the proof directory contains\n`events.ndjson`, `summary.json`, `gateway.log`, and\n`mock-openai-requests.ndjson`.\n\n## 5. Judge the evidence\n\nStart at the sent action in `summary.json`. Judge only later events from the\nselected SUT. Use stable `botApiMessageId` values to connect messages, edits,\nreactions, and deletions. Require a provider request when the path should reach\nthe model; native commands can correctly produce none. Read the\n[event model](features\u002Fruntime-reference.md#evidence-model) only when normalized\nfields need interpretation.\n\nReport the exact command with secrets omitted, the sent action and message id,\nthe relevant timeline rows, the provider request count, and the claim those\nfacts prove. A bare success marker or final screenshot is not sufficient.\n\nDone when every part of the claim has one Telegram-observable fact and every\nexpected model boundary has matching provider evidence.\n\n## Cleanup\n\nThe runner stops only its own process groups and removes credential and gateway\nscratch state. An explicit proof directory survives.\nKeep it until the review or repro no longer needs it.\n\nDone when no runner-owned process or listener remains, credential scratch state\nis gone, the Convex lease is released, assigned ports are free, and the named\nproof files remain readable.\n\nUse the dedicated QA user and Test Server bots. Keep every credential in its\nsource, active lease, or private runner scratch, and redact bot ids and\nusernames when evidence leaves the local task.\n","# Telegram E2E (Userbot)\n\nProve OpenClaw behavior against Telegram's Test Server as a dedicated QA user.\nThe user driver sees edits, deletions, reactions, and typing that a second bot\ncannot observe.\n\nTDLib is the headless user transport. Evidence is a structured Telegram event\ntimeline.\n\nEach credential contains one SUT bot and one independent TDLib authorization\nfor the shared QA user identity. Independent authorizations keep parallel runs\nfrom splitting Telegram updates between observers.\n\nPool creation, session repair, and credential publication remain owner-only\noperations outside this repository skill.\n\nConvex controls allocation only. During a lease, the worker may change the\ntest config, edit this harness, run arbitrary argv commands, and use the leased\nTDLib session or any Telegram Test Bot API method.\n\n## 1. Prepare\n\nRun from the OpenClaw checkout and ref under test. Point to the repository skill:\n\n```bash\nTELEGRAM_E2E_SKILL_DIR=\"${TELEGRAM_E2E_SKILL_DIR:-$PWD\u002F.agents\u002Fskills\u002Ftelegram-e2e-userbot}\"\nexport TELEGRAM_E2E_SKILL_DIR\n```\n\nThe runner starts the built `dist\u002Fentry.js` from this checkout. Before acquiring a\ncredential, verify that the exact ref has a dependency-ready runtime and usable\nbuild output without an implicit install or build. A bare worktree with missing\ndependencies or a missing or stale build stamp is not ready. Move to a trusted\ndependency-ready checkout or exact-head artifact first.\n\nOn hosts where builds are prohibited, `--source-gateway` runs a dependency-ready\nTypeScript checkout through the repository's existing development launcher.\n\nUse two unused ports for every run on a shared host:\n\n```bash\n: \"${TELEGRAM_GATEWAY_PORT:?set an unused Gateway port}\"\n: \"${TELEGRAM_MOCK_PORT:?set an unused provider port}\"\nexport TELEGRAM_GATEWAY_PORT TELEGRAM_MOCK_PORT\n```\n\nTeam and ClawSweeper runs use Convex. A maintainer with Convex CLI access to the\nOpenClaw broker project needs no local broker settings. The lease helper uses\n`qa\u002Fconvex-credential-broker` in the checkout, reads the production site and\nCI role through the authenticated CLI, and keeps them in process memory. Run\nthe same doctor and runner commands below; no credential export is required.\nRequire an authenticated `convex` command on `PATH`. When it is missing or\ncannot access the broker project, stop and ask the user to install and\nauthenticate the Convex CLI. Runtime setup must never install or log in.\n\nCI or another non-interactive worker can instead provide the broker pair:\n\n```bash\n: \"${OPENCLAW_QA_CONVEX_SITE_URL:?missing Convex QA site}\"\n: \"${OPENCLAW_QA_CONVEX_SECRET_CI:?missing Convex QA CI credential}\"\n```\n\nDone when the checkout contains `scripts\u002Fe2e\u002Fmock-openai-server.mjs`, the skill\ndirectory resolves, the exact runtime can start without setup work, and either\nthe Convex CLI can access the production broker project or the broker pair is\npresent.\n\n## 2. Select the proof\n\nRead [the verification map](features\u002FREADME.md), then read only the recipe for\nthe behavior under test. Prefer a DM for isolation. Use the shared group only\nwhen group policy, mentions, commands, topics, or reactions are part of the\nclaim.\n\nExercise the exact behavior changed by the diff. The generic `OPENCLAW_E2E_OK`\nturn proves only the default message path. Formatting, commands, media, edits,\ndeletions, reactions, topics, and timing each need actions and recorded events\nthat expose the specific claim.\n\nTreat the harness as an extension point. When its current actions, recorder\nfields, or recipes cannot expose the claim, extend them in the tested checkout.\nUse the leased TDLib session or any Telegram Test Bot API method needed to cover\nthe behavior.\n\nFor a non-default backend or timed scenario, read the matching section of the\n[runtime reference](features\u002Fruntime-reference.md).\n\nUse the scenario `command` action when the proof needs custom setup, inspection,\nor a Telegram API call. It receives the leased test bot, TDLib session, local\nTest Bot API proxy, Gateway config, and Gateway state.\n\nDone when the prompt, chat type, config patch, and expected event sequence each\nmap to the behavior being proved.\n\n## 3. Run the doctor\n\n```bash\nnode \"$TELEGRAM_E2E_SKILL_DIR\u002Fscripts\u002Ftelegram-test-doctor.mjs\"\n```\n\nRequire `ok: true`. The doctor leases one credential, restores its private TDLib\nstate, confirms the pinned TDLib version on the Test Server, verifies the SUT\nthrough the Test Bot API proxy, then removes the state and releases the lease.\n\nDone when the doctor reports Convex source, loaded credentials, isolated TDLib\nstate, Test Server, authorized user, Bot API proxy, disabled group privacy,\nactive group membership, and the SUT bot.\n\n## 4. Drive and record\n\nCreate a durable proof directory outside runner scratch state:\n\n```bash\nTELEGRAM_E2E_PROOF_DIR=\"$(mktemp -d \u002Ftmp\u002Ftelegram-e2e-proof.XXXXXX)\"\nnode \"$TELEGRAM_E2E_SKILL_DIR\u002Fscripts\u002Frun-mock-sut-user-e2e.mjs\" \\\n  --gateway-port \"$TELEGRAM_GATEWAY_PORT\" --mock-port \"$TELEGRAM_MOCK_PORT\" \\\n  --dm --text 'Please answer with OPENCLAW_E2E_OK only.' \\\n  --record \"$TELEGRAM_E2E_PROOF_DIR\u002Fevents.ndjson\" \\\n  --output \"$TELEGRAM_E2E_PROOF_DIR\u002Fsummary.json\"\n```\n\nThe runner owns one Convex lease, the Test Bot API proxy, a\nfresh gateway, the selected provider, the real-user action, recording, teardown,\nand cleanup. The default gateway and mock-provider ports are `19879` and `19882`;\nmanaged workloads on one host always pass distinct `--gateway-port` and\n`--mock-port` values instead of using defaults.\n\nConvex selects and exclusively leases any free credential. An exhausted pool\nwaits briefly for a release. The runner heartbeats during the run and releases\nthe credential during cleanup.\n\nRecording captures facts and rejects probe assertions such as `--expect` and\n`--any-sut-reply`. Use probe mode only for a quick reply check.\n\nDone when the runner exits zero and the proof directory contains\n`events.ndjson`, `summary.json`, `gateway.log`, and\n`mock-openai-requests.ndjson`.\n\n## 5. Judge the evidence\n\nStart at the sent action in `summary.json`. Judge only later events from the\nselected SUT. Use stable `botApiMessageId` values to connect messages, edits,\nreactions, and deletions. Require a provider request when the path should reach\nthe model; native commands can correctly produce none. Read the\n[event model](features\u002Fruntime-reference.md#evidence-model) only when normalized\nfields need interpretation.\n\nReport the exact command with secrets omitted, the sent action and message id,\nthe relevant timeline rows, the provider request count, and the claim those\nfacts prove. A bare success marker or final screenshot is not sufficient.\n\nDone when every part of the claim has one Telegram-observable fact and every\nexpected model boundary has matching provider evidence.\n\n## Cleanup\n\nThe runner stops only its own process groups and removes credential and gateway\nscratch state. An explicit proof directory survives.\nKeep it until the review or repro no longer needs it.\n\nDone when no runner-owned process or listener remains, credential scratch state\nis gone, the Convex lease is released, assigned ports are free, and the named\nproof files remain readable.\n\nUse the dedicated QA user and Test Server bots. Keep every credential in its\nsource, active lease, or private runner scratch, and redact bot ids and\nusernames when evidence leaves the local task.\n",{"name":5,"metadata":24,"description":7},{"argument-hint":25,"short-description":26},"\u003Cmessage-or-command?>","Telegram E2E via real-user driver","main","d69096e239f2a89aabc2cd078c47664a9d3b1e11","06-前端与设计\u002Ftelegram-e2e-userbot",3,"https:\u002F\u002Fgithub.com\u002Flza6\u002FArmor-piercing-system-prompts","https:\u002F\u002Fgithub.com\u002Flza6","个人vx：Tf00798\r\n个人✈：@TFGY999","2026-10-08 14:29:22",{"source":8,"skillName":5,"commands":36,"prompt":46,"zipAvailable":47},[37,40,43],{"tool":38,"command":39},"npx","npx skills add lza6\u002FArmor-piercing-system-prompts --skill telegram-e2e-userbot",{"tool":41,"command":42},"bunx","bunx skills add lza6\u002FArmor-piercing-system-prompts --skill telegram-e2e-userbot",{"tool":44,"command":45},"pnpm","pnpm dlx skills add lza6\u002FArmor-piercing-system-prompts --skill telegram-e2e-userbot","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Flza6\u002FArmor-piercing-system-prompts\u002Fblob\u002Fmain\u002F06-前端与设计\u002Ftelegram-e2e-userbot\u002FSKILL.md",true,[49,55,59,63,67,71],{"path":50,"resourceType":51,"resourceTypeDesc":52,"sizeBytes":53,"sourceHtmlUrl":54},"agents\u002Fopenai.yaml","REFERENCE","参考文档",351,"https:\u002F\u002Fgithub.com\u002Flza6\u002FArmor-piercing-system-prompts\u002Fblob\u002Fmain\u002F06-前端与设计\u002Ftelegram-e2e-userbot\u002Fagents\u002Fopenai.yaml",{"path":56,"resourceType":51,"resourceTypeDesc":52,"sizeBytes":57,"sourceHtmlUrl":58},"features\u002Fbasic-turns.md",2804,"https:\u002F\u002Fgithub.com\u002Flza6\u002FArmor-piercing-system-prompts\u002Fblob\u002Fmain\u002F06-前端与设计\u002Ftelegram-e2e-userbot\u002Ffeatures\u002Fbasic-turns.md",{"path":60,"resourceType":51,"resourceTypeDesc":52,"sizeBytes":61,"sourceHtmlUrl":62},"features\u002Fdelivery-lifecycle.md",2187,"https:\u002F\u002Fgithub.com\u002Flza6\u002FArmor-piercing-system-prompts\u002Fblob\u002Fmain\u002F06-前端与设计\u002Ftelegram-e2e-userbot\u002Ffeatures\u002Fdelivery-lifecycle.md",{"path":64,"resourceType":51,"resourceTypeDesc":52,"sizeBytes":65,"sourceHtmlUrl":66},"features\u002Freaction-lifecycle.md",2505,"https:\u002F\u002Fgithub.com\u002Flza6\u002FArmor-piercing-system-prompts\u002Fblob\u002Fmain\u002F06-前端与设计\u002Ftelegram-e2e-userbot\u002Ffeatures\u002Freaction-lifecycle.md",{"path":68,"resourceType":51,"resourceTypeDesc":52,"sizeBytes":69,"sourceHtmlUrl":70},"features\u002FREADME.md",5315,"https:\u002F\u002Fgithub.com\u002Flza6\u002FArmor-piercing-system-prompts\u002Fblob\u002Fmain\u002F06-前端与设计\u002Ftelegram-e2e-userbot\u002Ffeatures\u002FREADME.md",{"path":72,"resourceType":51,"resourceTypeDesc":52,"sizeBytes":73,"sourceHtmlUrl":74},"features\u002Fruntime-reference.md",6250,"https:\u002F\u002Fgithub.com\u002Flza6\u002FArmor-piercing-system-prompts\u002Fblob\u002Fmain\u002F06-前端与设计\u002Ftelegram-e2e-userbot\u002Ffeatures\u002Fruntime-reference.md",[76,83,88,95,102,109],{"skillNo":77,"name":78,"nameZh":6,"description":79,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":80,"categories":81,"keywords":82,"publishedAt":19},"01M7FFG485KY83TD6HSC4YPCMC","obsidian-clipper-template-creator","Guide for creating templates for the Obsidian Web Clipper. Use when you want to create a new clipping template, understand available variables, or format clipped content.","https:\u002F\u002Fgithub.com\u002Flza6\u002FArmor-piercing-system-prompts\u002Fblob\u002Fmain\u002F07-后端与架构\u002Fobsidian-clipper-template-creator\u002FSKILL.md",[],[],{"skillNo":84,"name":78,"nameZh":6,"description":79,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":85,"categories":86,"keywords":87,"publishedAt":19},"01M7FFG0T6AZ0STEJPSB3CX84K","https:\u002F\u002Fgithub.com\u002Flza6\u002FArmor-piercing-system-prompts\u002Fblob\u002Fmain\u002F07-后端与架构\u002Fobsidian-clipper-template-creator-antigravity-awesome-skills-main\u002FSKILL.md",[],[],{"skillNo":89,"name":90,"nameZh":6,"description":91,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":92,"categories":93,"keywords":94,"publishedAt":19},"01M7FFFY0QEM14SKWPDYGMXM2P","python-api","Programmatic Object-Detection-Metrics API guidance for BoundingBox, BoundingBoxes, Evaluator, enums, coordinate conversion, AP\u002FmAP semantics, and source-style import diagnostics.","https:\u002F\u002Fgithub.com\u002Flza6\u002FArmor-piercing-system-prompts\u002Fblob\u002Fmain\u002F07-后端与架构\u002Fobject-detection-metrics\u002Fsub-skills\u002Fpython-api\u002FSKILL.md",[],[],{"skillNo":96,"name":97,"nameZh":6,"description":98,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":99,"categories":100,"keywords":101,"publishedAt":19},"01M7FFFVZQZ2GJ0C0TGP7JHD20","file-evaluation","Folder-based PASCAL VOC AP\u002FmAP evaluation from ground-truth and detection text files.","https:\u002F\u002Fgithub.com\u002Flza6\u002FArmor-piercing-system-prompts\u002Fblob\u002Fmain\u002F07-后端与架构\u002Fobject-detection-metrics\u002Fsub-skills\u002Ffile-evaluation\u002FSKILL.md",[],[],{"skillNo":103,"name":104,"nameZh":6,"description":105,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":106,"categories":107,"keywords":108,"publishedAt":19},"01M7FFFSFNYSE77SJ6PE1CK2DA","object-detection-metrics","Guides legacy Object-Detection-Metrics PASCAL VOC AP\u002FmAP evaluation from detection text folders and direct Python API classes.","https:\u002F\u002Fgithub.com\u002Flza6\u002FArmor-piercing-system-prompts\u002Fblob\u002Fmain\u002F07-后端与架构\u002Fobject-detection-metrics\u002FSKILL.md",[],[],{"skillNo":110,"name":111,"nameZh":6,"description":112,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":113,"categories":114,"keywords":115,"publishedAt":19},"01M7FFFP0GCK1RVXBCPQ8APBMT","nx-workspace-patterns","Configure and optimize Nx monorepo workspaces. Use when setting up Nx, configuring project boundaries, optimizing build caching, or implementing affected commands.","https:\u002F\u002Fgithub.com\u002Flza6\u002FArmor-piercing-system-prompts\u002Fblob\u002Fmain\u002F07-后端与架构\u002Fnx-workspace-patterns\u002FSKILL.md",[],[]]