Agent Skills by Elliot
Showing 32 vetted skills indexed across 1 GitHub repositories.
ElliotJLTdrip
Quantify Claude interaction water footprints in milliliters using token counts.
ElliotJLTgranola-sync
Convert Granola cache meeting data into markdown notes with frontmatter.
ElliotJLTpipeline
Coordinate sequential design, implement, test, and review stages with checkpoints.
ElliotJLTmap-reduce
Execute map and reduce phases to aggregate batch processing results.
ElliotJLTstay-in-lane
Validate proposed code changes against user requests before editing.
ElliotJLTtrace-it
Traces callers and dependents before editing shared code to map impact.
ElliotJLTskill-forge
Generate structured SKILL.md files from session discoveries.
ElliotJLTpre-mortem
Generate failure scenarios and assess risk for project changes.
ElliotJLTyou-sure
Blocks dangerous operations until user confirms via a checklist.
ElliotJLTbattle-plan
Orchestrate a phased planning ritual for development work.
ElliotJLTcareful-delete
Require safety checks before destructive operations across file systems, databases, and version control.
ElliotJLTretrospective
Document post-task reflections with sections for failed attempts, what worked, and learnings.
ElliotJLTzero-in
Define search scope, expected results, and aliases before codebase exploration.
ElliotJLTpair-mode
Explain coding decisions and teach patterns during pair programming sessions.
ElliotJLTsafe-refactor
Orchestrate risk-aware refactor cycles with assessment, implementation, verification, and documentation phases.
ElliotJLTdont-be-greedy
Estimate token costs for CSV, JSON, and TXT files before loading.
ElliotJLTskill-creator
Create standardized SKILL.md templates with frontmatter, instructions, guardrails, and examples.
ElliotJLTlearn-from-this
Convert debugging failures into SKILL.md entries with required frontmatter.
ElliotJLTkeep-it-simple
Enforce YAGNI principles before creating factories, abstract base classes, or config-driven features.
ElliotJLTagent-audit
Identify load-bearing components in LLM agents via instrumentation and ablation testing.
ElliotJLTfan-out
Decompose tasks into independent subtasks and dispatch them to parallel agents.
ElliotJLTrubber-duck
Guide problem articulation through five structured questioning steps before proposing solutions.
ElliotJLTloose-ends
Detect unused imports, TODOs, and missing tests across changed files.
ElliotJLTprove-it
Run test suites and validate results before declaring completion.