timothymeyers
UserAgents, skills, and starter kits for Microsoft Discovery — the AI for Science platform.
Categories
Indexed Skills (14)
build-research-paper
Cross-domain / Scientific writing and publication — Build beautifully-formatted, arxiv-ready research paper PDFs from a mix of hand-authored LaTeX and/or markdown + PNG figures, using `tectonic` (default) or `pandoc` (escape hatch). Captures the reusable pattern that research projects keep re-discovering: `report/main.tex` + `references.bib` + `generate_figures.py` + iterative feedback/response scaffolding, all buildable with a single command. WHEN: write a research paper, arxiv submission, prepub, generate PDF from markdown, LaTeX paper, build main.tex, journal manuscript, referee revision, response to referee, resubmission package, publication-quality figures, matplotlib publication style, tectonic build, pandoc build, generate arxiv-ready PDF, bibliography from BibTeX.
discovery-token-usage
Cross-domain / Research platform observability — Mine a Discovery App workspace's .discovery/ exhaust for token usage across engines (Clio / Science Engine, copilot-cli / Mission Control / Generic Copilot), models, and interactive chat. Produces a console read-out by default; can also emit a markdown report on request. Cost / billing-multiplier data is a FEATURE FLAG — never surface it unless the user explicitly asks ("with cost", "include cost", "--with-cost", "premium units"). Handles seven data-quality traps (double-emitted stderr, corrupt journal appends, cumulative checkpoints, router duplicates, cached-read-in-realInput, etc). WHEN: "token usage", "how many tokens", "tokens in/out", "Discovery exhaust", "engine spend", "Clio vs copilot-cli", "mine .discovery/", "engine token read-out", "which model used the most tokens", "prompt inventory", "session_shutdown metrics", "TokenJournal".
osti-literature-search
Cross-domain / Scientific literature retrieval — Search and retrieve DOE-funded scientific literature from OSTI.GOV using `osti-axi`, a third-party read-only CLI. NOTE: this skill is a POINTER — the tool is NOT bundled with this catalog and must be installed separately from its own repository, with the user's explicit consent. Use for finding papers, reports, and datasets funded by the U.S. Department of Energy, pulling a record's abstract and resource links, and downloading OSTI-provided PDF/text full text into the workspace. WHEN: search OSTI, OSTI.GOV, DOE research papers, national laboratory publications, find DOE-funded literature, scientific literature search, download full text paper, get paper PDF, osti-axi, OSTI record, prior art for a DOE proposal, literature review for a national lab project.
alphafold-structure-provenance
Life sciences / Structural biology — Taxon-scoped UniProt accession lookup plus version-aware AlphaFold DB structure retrieval. Keeps pLDDT/PAE quality triage strictly separate from oligomer and complex assembly flags, and requires recording exactly which AlphaFold structure version any derived quantity (such as a HullRad diffusion coefficient) was computed from. Encodes AlphaFold DB version-retention behaviour: latestVersion is authoritative for what can be downloaded today, while allVersions is a historical list only — older version-numbered structure, PAE, and confidence files typically 404 once an entry moves to a newer version. WHEN: "look up UniProt accessions for an organism", "fetch AlphaFold structures", "AlphaFold API", "pLDDT triage", "PAE triage", "structure quality assessment", "AlphaFold model version", "latestVersion vs allVersions", "oligomer flag", "structure provenance", "AlphaFold DB 404".
biochemical-identifier-crosswalk
Life sciences / Biochemistry — Reconcile reactions, compounds, proteins, and taxa across Rhea, ChEBI (via EBI OLS4), MetaNetX, and UniProt into one one-to-many identifier crosswalk without collapsing chemical or mapping ambiguity. Config-driven: entities are declarative YAML records, so adding a reaction, compound, or enzyme means adding a record rather than editing code. Resolves reaction direction from the retrieved equation text rather than the Rhea LR/RL label, verifies mass and net-charge balance against an independently retrieved ChEBI formula and charge table, and never treats name similarity or a shared EC number alone as proof of identity. WHEN: "map Rhea to ChEBI, UniProt, or MetaNetX", "identifier crosswalk", "reaction identifier reconciliation", "resolve reaction direction", "preserve isoenzyme ambiguity", "is this the same reaction or compound", "mass and charge balance check", "currency metabolite vs generic compound".
biocyc-network-context
Life sciences / Systems biology — Parameterized bounded BioCyc, EcoCyc, and CyanoCyc network-context traversal: reaction -> enzyme/complex -> gene -> pathway. Preserves every catalyst for a reaction rather than collapsing isoenzymes, expands multi-subunit complexes to each subunit's own gene, and verifies organism identity from the PGDB's own NCBI-TAXONOMY-DB cross-reference rather than assuming it from the orgid string. Distinguishes missing_frame (the frame genuinely does not exist) from gated (a bot or subscription challenge blocked the page) instead of conflating both as "access failed". Rate-limited getxml client, never a bulk endpoint. Every identifier is a parameter. WHEN: "query BioCyc", "query EcoCyc", "query CyanoCyc", "getxml traversal", "BioCyc network context", "reaction to gene to pathway", "BioCyc isoenzyme", "protein complex subunits", "missing frame vs gated", "BioCyc rate limit", "BioCyc Limited Use License".
brenda-enzyme-data
Life sciences / Enzymology — Parameterized BRENDA ingestion from a caller-supplied licensed bulk JSON release. Subset extractor and deterministic canonical-record builder where organism substring, EC number, and file path are all parameters. Includes a BRENDA numeric-value-string parser covering ranges, the -999 {more} sentinel, and ligand-in-braces labels, plus a schema validator that surfaces a genuine publisher-side EC-key regex defect instead of hiding it behind a false "0 errors". You must supply your own licensed BRENDA download: this skill never bundles, caches, downloads, or accepts the license for any BRENDA dump. WHEN: "BRENDA", "brenda-enzymes.org", "BRENDA bulk JSON", "km_value", "turnover_number", "kcat_km_value", "specific_activity", "BRENDA -999 sentinel", "BRENDA schema validation", "BRENDA license", "ingest enzyme kinetic parameters from BRENDA".
discovery-add-agents
Cross-domain / Research platform engineering — Add agents from the public microsoft/discovery GitHub repository to a workspace, then deploy each selected agent to exactly one target: (a) the local Copilot app, emitting a `.github/agents/<name>.agent.md` companion invocable via `runSubagent`; (b) a Microsoft Discovery project plus Azure via `UpsertAgent`, binding existing tool ARM resources or building and registering missing ones; or (c) promoting an already-local agent up to a Discovery project without re-fetching upstream. USE FOR: 'add the discovery agents', 'download discovery agents', 'install all Microsoft discovery agents', 'add agents from microsoft/discovery', 'create copilot agents from microsoft/discovery', 'deploy discovery agents to Foundry', 'deploy discovery agents to Azure', 'promote a local agent to Microsoft Discovery'. DO NOT USE FOR: running a deployed agent (use runSubagent), or editing one agent's prompt (edit `agent.yaml`).
discovery-dashboard
Cross-domain / Research platform observability — Read-only live operations dashboard for a Microsoft Discovery workspace, built by scraping `.discovery/` and git. Shows leaf-task progress with approved-complete and awaiting-review kept strictly separate, the executing / ready / blocked work front from live `dependsOn` edges, engine liveness verified against real process identity, agent runs deduplicated across an alias graph, observed CLIO investigations, and a coverage panel separating "could not read" from "genuinely empty". Never writes to the workspace. WHEN: "project dashboard", "discovery dashboard", "what's the status of this project", "what's stuck", "how close are we to done", "are the engines still running", "is anything running", "are any agents running", "is clio running", "open the dashboard".
discovery-design-tasks
Cross-domain / Research methodology and experiment design — Generate a Microsoft Discovery task design (markdown only) for a user-specified local- app or cloud-service Foundry agent set using one (or all) of the three advanced investigation patterns: Deterministic, Guided Exploration, and Autonomous Research. For cloud-service task design or task creation, first look for a Discovery MCP server or Discovery proxy to resolve cloud workspaces, projects, agents, investigations, and tasks; if none is available, try Azure CLI before asking the user for another access route. Use when asked to plan a Discovery investigation, draft task trees, or choose an investigation pattern for an existing agent set.
sabio-rk-kinetics
Life sciences / Enzymology — Parameterized SABIO-RK kinetic-law ingestion. Organism, genus, and EC-number query tiering (exact organism -> genus wildcard -> higher-taxon lineage -> any organism) driven by taxid ancestry rather than lineage-name string matching. Explicit outcome and failure_kind semantics that never conflate a confirmed-empty result with a transport, authorization, or parser failure. Sample-vs-exhaustive pagination with honest truncation reporting, and Retry-After-aware handling of the documented 60-requests/60-seconds rate limit. Nothing is scoped to an organism, project, or EC number — every identifier is a parameter. WHEN: "query SABIO-RK", "SABIO-RK kinetic law", "SABIO-RK Export API", "ingest kinetic parameters", "kcat", "Km", "enzyme kinetics for an organism", "EC number query tiers", "organism lineage fallback", "taxid ancestry", "SABIO-RK pagination", "SABIO-RK 429", "SABIO-RK rate limit".
scientific-database-access
Cross-domain / Research data access and provenance — Establish a working, authorized, versioned route to a scientific database or data source before declaring it inaccessible. WHEN a user says "validate database access", "check dataset availability", "API returns HTML instead of JSON", "database needs credentials", "find an official bulk mirror", "is this source down or gated?", "the legacy API is dead", "404 vs paywall", "record access evidence", or asks whether a scientific database (SABIO-RK, BRENDA, BioCyc/CyanoCyc, Rhea, ChEBI, MetaNetX, UniProt, AlphaFold DB, or a similar source) is reachable/usable, use this skill before writing an adapter, before declaring a source "unavailable", and before designing any access-failure fallback.
scientific-evidence-qc
Cross-domain / Data quality and reproducibility — Produce an auditable, append-only observation layer and coverage/QC report over merged scientific-database evidence (kinetics, reaction stoichiometry, identifiers, provenance). WHEN a user says "normalize scientific observations", "merge database evidence", "audit provenance", "coverage and missingness report", "validate units", "QC the evidence layer", "is this traceability check real", "how many kinetic laws vs parameter rows", or asks you to build/extend a cross-source QC script over reaction-evidence-schema-shaped records, use this skill before writing new QC invariants, before trusting an existing "N/N passed" QC summary, and before treating a schema-conformance or requirement audit as sufficient evidence that the QC logic itself is correct.
discovery-project-story
Cross-domain / Research communication — Turn a Microsoft Discovery project's own record (git history, task board, engine/agent/Clio runs, outcomes and grades, the human side of chat sessions, token usage) into an evidence-true, animated, self-contained HTML presentation of how the work happened. The number of scenes and beats comes from the evidence and needs the user's approval before any prose is written. Each scene renders as a tree, a flow graph or a metrics chart, and every human quote is exact and re-verified at build time. WHEN: "tell the story of this project", "what happened in this project", "present how Discovery did this work", "project retrospective deck", "replay the project", "animated walkthrough of the work", "how was this built with Discovery", "project story HTML", "explain the project to leadership", "demo what Discovery did". DO NOT USE FOR: writing the research paper itself (arxiv-paper), PowerPoint decks (pptx), or token-only read-outs (discovery-token-usage).
Bio shown is the top-scored skill's repo description as a fallback — real GitHub bios land in a future update.