Updates a presentation with new numbers — quarterly refreshes, earnings updates, comp rolls, rebased market data. Use whenever the user asks to "update the deck with Q4 numbers", "refresh the comps", "roll this forward", "swap in the new earnings", "change all the $485M to $512M", or any request to swap figures across an existing deck without rebuilding it.
npx skills add https://github.com/anthropics/financial-services --skill deck-refresh
Update numbers across the deck. The deck is the source of truth for formatting; you're only changing values.
This skill works in both the PowerPoint add-in and chat. Identify which you're in before starting — the edit mechanism differs, the intent doesn't:
Either way: smallest possible change, existing formatting stays intact.
This is a four-phase process and the third phase is an approval gate. Don't edit until the user has seen the plan.
Use ask_user_question to find out how the new numbers are arriving:
Also ask about derived numbers: if revenue moves, does the user want growth rates and share percentages recalculated, or left alone? Most decks have "+15% YoY" baked in somewhere that's now stale. Whether to touch those is a judgment call the user should make, not you.
Read every slide. For each old value, find every instance — including the ones that don't look the same:
| Variant | Example |
|---|---|
| Scale | $485M, $0.485B, $485,000,000 |
| Precision | $485M, $485.0M, ~$485M |
| Unit style | $485M, $485MM, $485 million, 485M |
| Embedded | "revenue grew to $485M", "a $485M business", axis labels |
A deck that says $485M on slide 3, 485 on slide 8's chart axis, and $485.0 million in a footnote on slide 15 has three instances of the same number. Find-replace misses two of them. You shouldn't.
Where numbers hide:
Build a list: for each old value, every location it appears, the exact text it appears as, and what it'll become. This list is the plan.
This is a destructive operation on a deck someone spent time on. Show the full change list before editing a single thing. Format it so it's scannable:
$485M → $512M (Revenue)
Slide 3 — Title box: "Revenue grew to $485M"
Slide 8 — Chart axis label: "485"
Slide 15 — Footnote: "$485.0 million in FY24 revenue"
$120M → $135M (Adj. EBITDA)
Slide 3 — Table cell
Slide 11 — Body text: "$120M of Adj. EBITDA"
FLAGGED — possibly derived, not in your mapping:
Slide 3 — "+15% YoY" (growth rate — stale if base year didn't change?)
Slide 7 — "12% market share" (was this computed from $485M / market size?)
The flagged section matters. You're not just executing a find-replace — you're catching the second-order effects the user would've missed at 11pm. If the mapping says $485M → $512M and slide 3 also has +15% YoY right next to it, that growth rate is probably wrong now. Flag it; don't silently fix it, don't silently leave it.
Use ask_user_question for the approval: proceed as shown, proceed but skip the flagged items, or let them revise the mapping first.
For each change, make the smallest edit that accomplishes it. How that happens depends on your environment:
Either way, the standard is the same:
$485M is 14pt navy bold inside a sentence, $512M is 14pt navy bold inside the same sentence.Don't reformat anything you didn't need to touch. The deck's existing style is correct by definition; you're a surgeon, not a renovator.
After the last edit, report what actually happened:
Updated 11 values across 8 slides.
Changed:
[the list from Phase 3, now past-tense]
Still flagged — did NOT change:
Slide 3 — "+15% YoY" (derived; confirm separately)
Slide 7 — "12% market share"
Run standard visual verification checks on every edited slide. A number that got longer ($485M → $1,205M) might now overflow its text box or push a table column width. Catch it before the user does.
$MM and the user's mapping says $M, match the deck, not the mapping. Values change; style stays.Integration with protocols.io API for managing scientific protocols. This skill should be used when working with protocols.io to search, create, update, or publish protocols; manage protocol steps and materials; handle discussions and comments; organize workspaces; upload and manage files; or integrate protocols.io functionality into workflows. Applicable for protocol discovery, collaborative protocol development, experiment tracking, lab protocol management, and scientific documentation.
Analyzes job descriptions and generates tailored resumes that highlight relevant experience, skills, and achievements to maximize interview chances
Generate Excalidraw diagrams from natural language descriptions. Use when asked to "create a diagram", "make a flowchart", "visualize a process", "draw a system architecture", "create a mind map", or "generate an Excalidraw file". Supports flowcharts, relationship diagrams, mind maps, and system architecture diagrams. Outputs .excalidraw JSON files that can be opened directly in Excalidraw.
Build and distribute Expo development clients locally or via TestFlight
Use when you have a written implementation plan to execute in a separate session with review checkpoints
Data structure for annotated matrices in single-cell analysis. Use when working with .h5ad files or integrating with the scverse ecosystem. This is the data format skill—for analysis workflows use scanpy; for probabilistic models use scvi-tools; for population-scale queries use cellxgene-census.
Benchling R&D platform integration. Access registry (DNA, proteins), inventory, ELN entries, workflows via API, build Benchling Apps, query Data Warehouse, for lab data management automation.
Comprehensive molecular biology toolkit. Use for sequence manipulation, file parsing (FASTA/GenBank/PDB), phylogenetics, and programmatic NCBI/PubMed access (Bio.Entrez). Best for batch processing, custom bioinformatics pipelines, BLAST automation. For quick lookups use gget; for multi-service integration use bioservices.
Take anthropics/deck-refresh from the repository into ~/.claude/skills for personal
use, or into .claude/skills inside a project.
The agent identifies a skill by the name field in its header. Two skills with the
same name cannot sit side by side — one of them will be ignored.