575 skills · sorted by relevance
Checks factual claims in articles and reports against current sources or independently queried original data. Applies corrections, assigns one of six confidence levels to each claim, and produces a sourced report with an overall pass or fail assessment.
Turns user corrections and recurring mistakes into actionable rules with confidence scores. It records when each rule applies and why it matters, so relevant lessons can be retrieved before similar work begins.
Build three-tier bookings forecasts from sales opportunities and historical conversion data. Project recurring revenue by customer cohort, assess the reliability of each sales stage, and present forecast numbers alongside their assumptions.
Estimate development effort using affected code, design documents, dependencies, and past sprint results. Get optimistic, expected, and pessimistic day estimates, along with a confidence level, key risks, and a suggested breakdown into smaller tasks.
Find open issues affected by a proposed code change. Identify additional problems it may fix or requested behavior it may conflict with, then report supporting code differences, issue symptoms, and confidence levels.
Turn selected research conversations or notes into themes supported by evidence and counterexamples. Distinguish observed behavior from preferences and hypotheses, assess confidence, and suggest a small next experiment with an observable measure of success.
Research topics across multiple sources, cross-check findings, and verify source URLs before producing a report with confidence labels. Choose from quick lookups to iterative investigations that preserve findings, with community sources prioritized for questions about public reactions.
Design systems that use language models to judge AI responses. Covers scoring criteria, comparisons between two answers, bias reduction, confidence calibration, and checks against human judgments to make automated quality assessments more consistent and traceable.
Convert PDF, PowerPoint, Excel, and Word documents into structured Markdown for search and information extraction. Outputs include document metadata, a parsing confidence score, and the parser used, with fallback parsers available when earlier attempts fail.
Classify messages, rate content, or estimate probabilities using defined question types. Receive machine-readable answers, with confidence scores for classification and ratings, and apply the same questions to multiple inputs in a batch.
Continue suitable tasks when web search repeatedly fails by assessing whether existing knowledge is sufficient. Mark unverified claims, distinguish confidence levels, and provide a follow-up checklist so readers can see what still needs checking.
Scan source code for security weaknesses with Semgrep after approving the target and rulesets. Choose full results or focus on important findings, then receive merged results and a record of failed or skipped scans.
Design charts and dashboards that show data distributions, uncertainty, missing values, and sample sizes clearly. Choose visual forms that preserve meaningful variation and explain where averages, aggregation, or interval choices could mislead readers.
Screen an existing compound library against a protein target with Boltz. Rank candidates by binding confidence or lead optimization score, download predicted structures, and review solubility, permeability, and lipophilicity results alongside the ranked hits.
Screen supplied proteins, peptides, antibodies, or nanobodies against a target with Boltz. Prepares inputs, obtains a cost estimate for approval, downloads results, and ranks candidates by binding confidence, a score for predicted binding reliability.
Generate new protein, peptide, antibody, or nanobody candidates that bind a chosen target using Boltz. The workflow covers target exploration, design setup, cost estimates, job submission, result downloads, and ranking by predicted binding confidence.
Generate new small molecules intended to bind a target protein with Boltz. The workflow covers cost estimation, job submission, result downloads, and ranking candidates by binding confidence or a score for improving promising molecules.
Use a panel of three to five agents to decide between options or assess a plan. Independent initial opinions and anonymous deliberation lead to one verdict with confidence, preserved dissent, a validation test, and a next action.
Turn a large codebase spanning multiple repositories or services into a structured knowledge base. It documents architecture, business rules, and dependencies, links conclusions to source lines, labels confidence in relationships, and supports incremental updates.
Extract lessons from corrections, praise, and unexpected cases in conversations. Organize findings by confidence level and propose updates to individual agent histories or shared team decisions, with changes shown for approval before saving.
Predict structures of complexes containing proteins, DNA, RNA, and ligands with OpenFold3, an AlphaFold3 reproduction with publicly available model weights. Prepare JSON inputs, generate candidate structures, and inspect confidence scores and molecular clashes.
Support authorized external security reconnaissance, the discovery of exposed systems and information. The skill provides search queries, probe references, secret detection patterns, and scoring guidance to help identify assets and organize findings with evidence and confidence levels.
Investigate where or when a photo was taken using visible clues, metadata, reverse image search, maps, and image comparisons. Results include coordinates, an error radius, evidence images, and confidence levels that distinguish verified findings from uncertain candidates.
Predict protein structures by connecting MSA-Search with OpenFold3. It finds related sequences, aligns them to provide evolutionary context, and saves predicted structures with confidence scores. It also supports comparing predictions with and without that context.