[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M7542582HWCWWVM50PDX1M4J":3,"skill-same:01M7542582HWCWWVM50PDX1M4J":50},{"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":28,"commitSha":29,"skillDirectory":30,"forksCount":31,"repositoryHtmlUrl":32,"repositoryHomepageUrl":33,"ownerHtmlUrl":34,"ownerBlogUrl":6,"ownerDescription":35,"updatedAt":36,"install":37,"resources":49},"01M7542582HWCWWVM50PDX1M4J","genomi-decode",null,"Activate this skill for \"\u002Fgenomi decode\", \"decode my genome\", \"decode my\nDNA\", \"show me the dashboard\", \"the Genomi dashboard\",\n\"one-shot rundown\", or any all-at-once request that asks Genomi to\ncompose every capability's findings into a single artifact. This is the\nwhole-genome dashboard kicker — it sweeps every relevant Genomi capability\nin one shot, not a per-target lookup.\n\nComposes evidence from every relevant Genomi capability into a single\nself-contained Genomi Dashboard.html and returns localhost serve metadata.\nActive genome required.","exon-research\u002Fgenomi","exon-research","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F287072540?v=4","01M754130SK0JVWXNQKF8XANKV","Exon Research","01M754131W4AFK5CDRWT1KHP8X",492,"Apache-2.0","https:\u002F\u002Fgithub.com\u002Fexon-research\u002Fgenomi\u002Fblob\u002Fmaster\u002Fskills\u002Fdecode\u002FSKILL.md",[],[],"2026-11-12 13:55:43",false,"---\nname: genomi-decode\ndescription: |\n  Activate this skill for \"\u002Fgenomi decode\", \"decode my genome\", \"decode my\n  DNA\", \"show me the dashboard\", \"the Genomi dashboard\",\n  \"one-shot rundown\", or any all-at-once request that asks Genomi to\n  compose every capability's findings into a single artifact. This is the\n  whole-genome dashboard kicker — it sweeps every relevant Genomi capability\n  in one shot, not a per-target lookup.\n\n  Composes evidence from every relevant Genomi capability into a single\n  self-contained Genomi Dashboard.html and returns localhost serve metadata.\n  Active genome required.\ntools:\n  - decode.render_dashboard\nmutating: true\n---\n\n# Genomi Decode\n\nThe `\u002Fgenomi decode` kicker tells the agent to assemble every relevant Genomi\ncapability's evidence about the user's active genome and emit a single\nself-contained `Genomi Dashboard.html` artifact. Activate this skill whenever\nthe user types `\u002Fgenomi decode`, asks for \"the dashboard\", asks to \"decode my\ngenome\", or asks for a one-shot evidence rundown.\n\n## Activation\n\nThis skill requires an Active Genome Index session and explicit approval to\nread it. The same approval gate that protects `variant.resolve`, `clinvar.*`,\nand the PGx ops protects `decode.render_dashboard`. If no active genome is\nselected the op fails with `active_genome_index_required`; if approval has not\nbeen granted it fails with `active_genome_index_approval_required`.\n\n## Reconcile Active Genome Index lifecycle before gathering panels\n\nCall `genomi.describe_context` first. If `active_genome_index.active_genome_index_readiness.status`\nis `needs_reparse` or `schema_too_new`, **handle the lifecycle before\ngathering any panel evidence** — do not proceed with a stale Active Genome Index and\nsilently bound the panels.\n\nThe full procedure lives in the Active Genome Index skill under the lifecycle\nguidance for `needs_reparse` and `schema_too_new`.\nSummary for decode:\n\n1. If `needs_reparse` and `availability.agi_intake_source_path` is true, call\n   `genomi.parse_source({\"source\": active_genome_index.agi_intake_source_path})` without\n   prompting. Routine maintenance.\n2. If `needs_reparse` and the source path is gone, ask the user once for\n   the current path and parse that. Don't continue with a stale Active Genome Index.\n3. If `schema_too_new`, the user's runtime is out of date — tell them to\n   upgrade Genomi, stop.\n4. Only after `active_genome_index_readiness.status == \"complete\"` call the\n   decode operation.\n\n## Dashboard Build\n\nCall `decode.render_dashboard`. Decode owns panel gathering, panel shaping,\nand rendering. The agent may choose dashboard categories through structured\nparameters such as `panels` and select declared score\u002Fdomain options. Omitted\n`panels` means every dashboard category. The agent does not assemble panel\nevidence and does not ask which PGx route to run; decode owns that work.\n\nThe renderer normalizes native upstream-op shapes internally:\n\n- `overview` — adapts `active_genome_index.summarize` output;\n  snake_case keys (`genome_build`, `nickname`, `active_genome_index_completed_at`,\n  `nearest_reference_groups`) are mapped automatically.\n- `variants` — adapts `clinvar.scan_candidates` variant inventory rows;\n  `clinvar.match_variants` JSONL rows (`{sample_variant, clinvar}`) are also\n  handled. Carrier\u002Fcondition review groups render under `risk`, not variants.\n- `nutrigenomics` — adapts `nutrigenomics.retrieve_domain_markers`; it extracts\n  `gene.symbol`, `variant.rsid`, `established_effect.claim` (→ `recommendation`),\n  `evidence_tier`, and domain label (→ `marker`).\n- `ancestry` — adapts `ancestry.estimate_population_context`.\n- `pgx` — adapts PharmCAT `sample_pgx_matrix` and medication-review\n  `medication_review_matrix` rows into PGx cards without merging separate\n  medication recommendations by gene alone.\n- `risk` — adapts native `prs.calculate_score` results and\n  `phenotype.plan_risk_investigation` carrier\u002Fcondition review rows into\n  risk\u002Freview cards.\n- `variants_all` — uses the ClinVar matches JSONL path materialized by decode.\n\nDecode also gathers the current carrier\u002Fcondition and PGx review contracts:\n\n- For `risk`, decode runs the declared `risk_review_types` from the selected\n  Active Genome Index ClinVar matches scope. Omitted `risk_review_types` means\n  `carrier_review` plus `observed_condition_review`; pass an empty array only\n  when the user wants PRS-only risk evidence.\n- For `pgx`, decode runs `pharmacogenomics.review_medication` for explicit\n  `pgx_review_targets` and for drug\u002Fgene targets discovered in PharmCAT\n  `sample_pgx_matrix` rows, up to `pgx_review_target_limit`. Gene-only sample\n  rows can be preserved as sample evidence, but decode does not invent\n  medication-specific recommendations without a declared drug\u002Fsource target.\n\nIf no PRS scores are installed in the user's library, the builder supplies a\ntyped empty risk state so stale risk evidence is cleared rather than preserved.\n\n## Verify before claiming success\n\nThe renderer's response is the source of truth:\n\n- `panels_rendered`: panels that landed with real data.\n- `panels_empty`: panels with no usable evidence — they render as\n  category-specific unavailable states in the UI.\n- `evidence_build.panels_running`: panels still running in a background job.\n- `evidence_build.panel_states`: per-panel source status, including PGx\n  background job ids and check operations when applicable.\n\nRead `panels_empty` and any `evidence_build.panel_states` before telling the\nuser the dashboard is ready. Surface incomplete categories honestly with their\ntyped state.\n\n## Refresh vs. reuse\n\nCall `decode.render_dashboard` again to refresh the dashboard after installing\nlibraries or changing category selections. Panels without usable evidence render\nas category-specific unavailable states.\n\n## Output location\n\nBy default the artifact is written to\n`\u003Ctmp>\u002Fgenomi-dashboards\u002F\u003Csample>\u002Fdashboard.html`. The user may override\n`output` with any absolute filesystem path; the parent directory is created on\ndemand.\n\n## Serving the dashboard\n\n`decode.render_dashboard` returns a `serve` block:\n\n```json\n{\n  \"serve\": {\n    \"status\": \"started\",\n    \"directory\": \"...\",\n    \"filename\": \"dashboard.html\",\n    \"port\": 8766,\n    \"url\": \"http:\u002F\u002F127.0.0.1:8766\u002Fdashboard.html\",\n    \"command\": \"python3 -m http.server 8766 --bind 127.0.0.1 --directory ...\"\n  }\n}\n```\n\nNormal runtime calls start a local static dashboard server automatically and\nchoose a free localhost port. Tell the user `serve.url`. If `serve.status` is\n`ready_to_start` or `start_failed`, run `serve.command` as a fallback and then\ntell the user the adjusted URL.\n\n## Boundaries\n\n- Active Genome Index session approval is required.\n- Decode owns panel evidence collection and shaping for the dashboard artifact.\n- The artifact is a single self-contained HTML file that renders fully offline\n  — React\u002FReactDOM and the precompiled app JS are inlined, no CDN, no\n  in-browser Babel. (One optional Google Fonts stylesheet is referenced; it\n  falls back to system fonts offline and carries no genome data.) It opens by\n  double-click; the local server is only there so the user can hit a URL.\n\n## Tool\n\n### decode.build_dashboard_evidence\n\nSupport operation used by `decode.render_dashboard` to inspect panel readiness\nand gaps. Normal dashboard requests should call `decode.render_dashboard`.\n\n### decode.render_dashboard\n\nBuild, shape, and render the Genomi Dashboard HTML artifact from the approved\nActive Genome Index. Returns\n`{ status, dashboard_path, panels_rendered, panels_empty, serve }` plus the\nstandard `evidence_envelope`. The `serve` block tells the host agent how to\nexpose the dashboard at a localhost URL — see the \"Serving the dashboard\"\nsection above.\n\n## Cross-Capability Synthesis\n\nA scope-limited result from this capability is not a final user-facing answer\nwhen other Genomi capabilities can contribute orthogonal evidence to the same\nquestion. Returning \"cannot answer\" while applicable capabilities remain\nunexamined is a host-agent failure mode.\n","# Genomi Decode\n\nThe `\u002Fgenomi decode` kicker tells the agent to assemble every relevant Genomi\ncapability's evidence about the user's active genome and emit a single\nself-contained `Genomi Dashboard.html` artifact. Activate this skill whenever\nthe user types `\u002Fgenomi decode`, asks for \"the dashboard\", asks to \"decode my\ngenome\", or asks for a one-shot evidence rundown.\n\n## Activation\n\nThis skill requires an Active Genome Index session and explicit approval to\nread it. The same approval gate that protects `variant.resolve`, `clinvar.*`,\nand the PGx ops protects `decode.render_dashboard`. If no active genome is\nselected the op fails with `active_genome_index_required`; if approval has not\nbeen granted it fails with `active_genome_index_approval_required`.\n\n## Reconcile Active Genome Index lifecycle before gathering panels\n\nCall `genomi.describe_context` first. If `active_genome_index.active_genome_index_readiness.status`\nis `needs_reparse` or `schema_too_new`, **handle the lifecycle before\ngathering any panel evidence** — do not proceed with a stale Active Genome Index and\nsilently bound the panels.\n\nThe full procedure lives in the Active Genome Index skill under the lifecycle\nguidance for `needs_reparse` and `schema_too_new`.\nSummary for decode:\n\n1. If `needs_reparse` and `availability.agi_intake_source_path` is true, call\n   `genomi.parse_source({\"source\": active_genome_index.agi_intake_source_path})` without\n   prompting. Routine maintenance.\n2. If `needs_reparse` and the source path is gone, ask the user once for\n   the current path and parse that. Don't continue with a stale Active Genome Index.\n3. If `schema_too_new`, the user's runtime is out of date — tell them to\n   upgrade Genomi, stop.\n4. Only after `active_genome_index_readiness.status == \"complete\"` call the\n   decode operation.\n\n## Dashboard Build\n\nCall `decode.render_dashboard`. Decode owns panel gathering, panel shaping,\nand rendering. The agent may choose dashboard categories through structured\nparameters such as `panels` and select declared score\u002Fdomain options. Omitted\n`panels` means every dashboard category. The agent does not assemble panel\nevidence and does not ask which PGx route to run; decode owns that work.\n\nThe renderer normalizes native upstream-op shapes internally:\n\n- `overview` — adapts `active_genome_index.summarize` output;\n  snake_case keys (`genome_build`, `nickname`, `active_genome_index_completed_at`,\n  `nearest_reference_groups`) are mapped automatically.\n- `variants` — adapts `clinvar.scan_candidates` variant inventory rows;\n  `clinvar.match_variants` JSONL rows (`{sample_variant, clinvar}`) are also\n  handled. Carrier\u002Fcondition review groups render under `risk`, not variants.\n- `nutrigenomics` — adapts `nutrigenomics.retrieve_domain_markers`; it extracts\n  `gene.symbol`, `variant.rsid`, `established_effect.claim` (→ `recommendation`),\n  `evidence_tier`, and domain label (→ `marker`).\n- `ancestry` — adapts `ancestry.estimate_population_context`.\n- `pgx` — adapts PharmCAT `sample_pgx_matrix` and medication-review\n  `medication_review_matrix` rows into PGx cards without merging separate\n  medication recommendations by gene alone.\n- `risk` — adapts native `prs.calculate_score` results and\n  `phenotype.plan_risk_investigation` carrier\u002Fcondition review rows into\n  risk\u002Freview cards.\n- `variants_all` — uses the ClinVar matches JSONL path materialized by decode.\n\nDecode also gathers the current carrier\u002Fcondition and PGx review contracts:\n\n- For `risk`, decode runs the declared `risk_review_types` from the selected\n  Active Genome Index ClinVar matches scope. Omitted `risk_review_types` means\n  `carrier_review` plus `observed_condition_review`; pass an empty array only\n  when the user wants PRS-only risk evidence.\n- For `pgx`, decode runs `pharmacogenomics.review_medication` for explicit\n  `pgx_review_targets` and for drug\u002Fgene targets discovered in PharmCAT\n  `sample_pgx_matrix` rows, up to `pgx_review_target_limit`. Gene-only sample\n  rows can be preserved as sample evidence, but decode does not invent\n  medication-specific recommendations without a declared drug\u002Fsource target.\n\nIf no PRS scores are installed in the user's library, the builder supplies a\ntyped empty risk state so stale risk evidence is cleared rather than preserved.\n\n## Verify before claiming success\n\nThe renderer's response is the source of truth:\n\n- `panels_rendered`: panels that landed with real data.\n- `panels_empty`: panels with no usable evidence — they render as\n  category-specific unavailable states in the UI.\n- `evidence_build.panels_running`: panels still running in a background job.\n- `evidence_build.panel_states`: per-panel source status, including PGx\n  background job ids and check operations when applicable.\n\nRead `panels_empty` and any `evidence_build.panel_states` before telling the\nuser the dashboard is ready. Surface incomplete categories honestly with their\ntyped state.\n\n## Refresh vs. reuse\n\nCall `decode.render_dashboard` again to refresh the dashboard after installing\nlibraries or changing category selections. Panels without usable evidence render\nas category-specific unavailable states.\n\n## Output location\n\nBy default the artifact is written to\n`\u003Ctmp>\u002Fgenomi-dashboards\u002F\u003Csample>\u002Fdashboard.html`. The user may override\n`output` with any absolute filesystem path; the parent directory is created on\ndemand.\n\n## Serving the dashboard\n\n`decode.render_dashboard` returns a `serve` block:\n\n```json\n{\n  \"serve\": {\n    \"status\": \"started\",\n    \"directory\": \"...\",\n    \"filename\": \"dashboard.html\",\n    \"port\": 8766,\n    \"url\": \"http:\u002F\u002F127.0.0.1:8766\u002Fdashboard.html\",\n    \"command\": \"python3 -m http.server 8766 --bind 127.0.0.1 --directory ...\"\n  }\n}\n```\n\nNormal runtime calls start a local static dashboard server automatically and\nchoose a free localhost port. Tell the user `serve.url`. If `serve.status` is\n`ready_to_start` or `start_failed`, run `serve.command` as a fallback and then\ntell the user the adjusted URL.\n\n## Boundaries\n\n- Active Genome Index session approval is required.\n- Decode owns panel evidence collection and shaping for the dashboard artifact.\n- The artifact is a single self-contained HTML file that renders fully offline\n  — React\u002FReactDOM and the precompiled app JS are inlined, no CDN, no\n  in-browser Babel. (One optional Google Fonts stylesheet is referenced; it\n  falls back to system fonts offline and carries no genome data.) It opens by\n  double-click; the local server is only there so the user can hit a URL.\n\n## Tool\n\n### decode.build_dashboard_evidence\n\nSupport operation used by `decode.render_dashboard` to inspect panel readiness\nand gaps. Normal dashboard requests should call `decode.render_dashboard`.\n\n### decode.render_dashboard\n\nBuild, shape, and render the Genomi Dashboard HTML artifact from the approved\nActive Genome Index. Returns\n`{ status, dashboard_path, panels_rendered, panels_empty, serve }` plus the\nstandard `evidence_envelope`. The `serve` block tells the host agent how to\nexpose the dashboard at a localhost URL — see the \"Serving the dashboard\"\nsection above.\n\n## Cross-Capability Synthesis\n\nA scope-limited result from this capability is not a final user-facing answer\nwhen other Genomi capabilities can contribute orthogonal evidence to the same\nquestion. Returning \"cannot answer\" while applicable capabilities remain\nunexamined is a host-agent failure mode.\n",{"name":5,"tools":24,"mutating":26,"description":27},[25],"decode.render_dashboard",true,"Activate this skill for \"\u002Fgenomi decode\", \"decode my genome\", \"decode my\nDNA\", \"show me the dashboard\", \"the Genomi dashboard\",\n\"one-shot rundown\", or any all-at-once request that asks Genomi to\ncompose every capability's findings into a single artifact. This is the\nwhole-genome dashboard kicker — it sweeps every relevant Genomi capability\nin one shot, not a per-target lookup.\n\nComposes evidence from every relevant Genomi capability into a single\nself-contained Genomi Dashboard.html and returns localhost serve metadata.\nActive genome required.\n","master","1df4f5bfb635e2d08f93a1102b70b840fbe80d86","skills\u002Fdecode",61,"https:\u002F\u002Fgithub.com\u002Fexon-research\u002Fgenomi","https:\u002F\u002Fwww.genomiagent.com\u002F","https:\u002F\u002Fgithub.com\u002Fexon-research","Building local-first AI co-scientist for genomic research","2026-09-01 03:30:56",{"source":8,"skillName":5,"commands":38,"prompt":48,"zipAvailable":26},[39,42,45],{"tool":40,"command":41},"npx","npx skills add exon-research\u002Fgenomi --skill genomi-decode",{"tool":43,"command":44},"bunx","bunx skills add exon-research\u002Fgenomi --skill genomi-decode",{"tool":46,"command":47},"pnpm","pnpm dlx skills add exon-research\u002Fgenomi --skill genomi-decode","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Fexon-research\u002Fgenomi\u002Fblob\u002Fmaster\u002Fskills\u002Fdecode\u002FSKILL.md",[],[51,58,65,72,79,86],{"skillNo":52,"name":53,"nameZh":6,"description":54,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":55,"categories":56,"keywords":57,"publishedAt":19},"01M75442YT3ZG6MG47E5MJ3HG2","variant-evidence","Answer specific rsID, allele, gene, region, genotype, and absence\u002Fcallability\nquestions using explicit session context or public evidence.","https:\u002F\u002Fgithub.com\u002Fexon-research\u002Fgenomi\u002Fblob\u002Fmaster\u002Fskills\u002Fvariant-evidence\u002FSKILL.md",[],[],{"skillNo":59,"name":60,"nameZh":6,"description":61,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":62,"categories":63,"keywords":64,"publishedAt":19},"01M754400MW6S0MEN19S4CM9DE","journal-source-research","Journal sub-skill for focused public\u002Fsource evidence review and reviewed\nfinding write-back before interpretation or answer synthesis.","https:\u002F\u002Fgithub.com\u002Fexon-research\u002Fgenomi\u002Fblob\u002Fmaster\u002Fskills\u002Fsource-research\u002FSKILL.md",[],[],{"skillNo":66,"name":67,"nameZh":6,"description":68,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":69,"categories":70,"keywords":71,"publishedAt":19},"01M7543X67TA0MBK3YGC0NJAX9","sequence","Deterministic sequence utilities for translation, ORFs, restriction sites,\nKozak context, primer checks, and local FASTA record matching.","https:\u002F\u002Fgithub.com\u002Fexon-research\u002Fgenomi\u002Fblob\u002Fmaster\u002Fskills\u002Fsequence\u002FSKILL.md",[],[],{"skillNo":73,"name":74,"nameZh":6,"description":75,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":76,"categories":77,"keywords":78,"publishedAt":19},"01M7543RHCHF4WJ69P0K0QTTFB","rare-disease-cancer","Plan rare disease, hereditary disease, cancer risk, carrier-relevance, and\nobserved-condition source investigation from public targets or selected\nactive genome evidence.","https:\u002F\u002Fgithub.com\u002Fexon-research\u002Fgenomi\u002Fblob\u002Fmaster\u002Fskills\u002Frare-disease-cancer\u002FSKILL.md",[],[],{"skillNo":80,"name":81,"nameZh":6,"description":82,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":83,"categories":84,"keywords":85,"publishedAt":19},"01M7543NDTVZB8T3RQBD0RK3MC","prs","Apply published polygenic scores from PGS Catalog to approved local personal DNA and return raw weighted score plus overlap QC.","https:\u002F\u002Fgithub.com\u002Fexon-research\u002Fgenomi\u002Fblob\u002Fmaster\u002Fskills\u002Fprs\u002FSKILL.md",[],[],{"skillNo":87,"name":88,"nameZh":6,"description":89,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":90,"categories":91,"keywords":92,"publishedAt":19},"01M7543FDFTWSYVVTEGHQSPHQK","pharmacogenomics","Answer drug-response, medication, PharmGKB-style, PGxDB, ATC, DrugBank,\ngene-drug, and variant-drug questions using public PGx evidence plus local\nsample genotype support when an Active Genome Index is selected.","https:\u002F\u002Fgithub.com\u002Fexon-research\u002Fgenomi\u002Fblob\u002Fmaster\u002Fskills\u002Fpharmacogenomics\u002FSKILL.md",[],[]]