daniel-ospina
@daniel-ospina
Public Skills
how-to-use-tortoise
by daniel-ospina
Create points, operators, mitigations, NANDs, supersede, delete, and annotate on the Tortoise graph. Use when asked to write to the Tortoise graph — creating points, operators, mitigations, NAND edges, superseding points, deleting, or annotating.
tortoise-file-finding
by daniel-ospina
Ingest a research finding into the Tortoise graph. Creates a Point, checks for existing related claims, and surfaces connection candidates.
tortoise-onboarding
by daniel-ospina
"Install and connect Tortoise for your agent — self-adjudicate your harness (Claude Code / Cursor / Codex / Pi = self-install; Claude Desktop / Claude Web = teach-human), write the MCP config, verify with tortoise_health, checkpoint harness-connected, and reach the decide protocol without any local skill file. Self-hosted: Docker-first install (Compose daemon + FalkorDB sidecar); embedded is the eval-only fallback. Successor to AGENT_ONBOARDING.md (archived)."
tortoise-onboarding
by daniel-ospina
"Install and connect Tortoise for your agent — self-adjudicate your harness (Claude Code / Cursor / Codex / Pi = self-install; Claude Desktop / Claude Web = teach-human), write the MCP config, verify with tortoise_health, checkpoint harness-connected, and reach the decide protocol without any local skill file. Self-hosted: Docker-first install (Compose daemon + FalkorDB sidecar); embedded is the eval-only fallback. Successor to AGENT_ONBOARDING.md (archived)."
tortoise-decide
by daniel-ospina
Make a decision with the Tortoise graph — refine the decision, research options/criteria/findings, wire IMPL/NAND/mitigation edges, and rank options by EP confidence. Invoke this skill when the user needs to decide between options and the graph should hold the reasoning.
code-review
by daniel-ospina
MANDATORY code review for standard+complex PRs. Dispatches parallel reviewers. Skipping ships unreviewed code to production.
commit-workflow
by daniel-ospina
MANDATORY before any git commit, push, or merge. Runs pre-flight (typecheck, tests, pgTAP), creates PR, code-review gate, auto-merge. Skipping bypasses ALL quality gates.
competitor-research
by daniel-ospina
Use when asked to "research a competitor", "analyze competitor X", "profile competitor", "competitive analysis", "competitor deep dive", or when evaluating a competitor's product, strategy, positioning, traction, or business model. Produces structured profiles following the skill's 11-dimension competitor template (bootstrapped to the output directory as _template.md on first run) and updates the competitive analysis synthesis.
clarifying-questions
by daniel-ospina
"Scores and surfaces clarifying questions before planning proceeds. Callers: issue-scoping (Phase 0.5, mode=issue-pre) and strategy-builder (§2.4, GTM domain). Not invoked directly by users."
carousel-designer
by daniel-ospina
Visual design critique for rendered carousel slides. Single Opus session critiques composite slide images, generates CSS fixes, and re-reviews in a continuous loop until clean. Final verification by separate DeepSeek model. Plugs into art-director Gate 2 for post-render carousel QA.
carousel-b2b-strategy
by daniel-ospina
7-phase stateful pipeline for B2B carousel content with status.yaml tracking and cross-session resume. Use when asked to create a carousel, social media post, or thought leadership content. NEVER create carousel content without reading this skill — this is a pipeline with persistent state, not a one-shot content task.
art-director
by daniel-ospina
Thin design system orchestrator for the carousel pipeline. Receives carousel content from carousel-b2b-strategy, enforces design rules, validates tokens and safe zones pre-render, and routes post-render output to QA skills. Not a design tool — routes and gates.
codebase-audit
by daniel-ospina
"Full codebase audit coordinating security, bug, config, supply-chain, and database specialists. Use when asked to 'audit the codebase', 'find all bugs', 'security audit', 'full audit', 'code quality scan', or 'run codebase audit'."
define-strategy
by daniel-ospina
"Defines organizational strategy — gap analysis, initiatives, role alignment, and roadmap — for any function, team, or role."
app-test
by daniel-ospina
"Automated E2E testing for desktop (Electron) and web apps. Launches app, runs test scenarios, screenshots, verifies logs and database state. Invoke when asked to test the app, verify a release, or smoke-test after deploy."
android-deploy
by daniel-ospina
Build and deploy the El Dato Scanner app to Google Play Internal Testing. Use when asked to "deploy the scanner", "build the APK/AAB", "upload to Play Store", or "ship a new scanner version".
carousel-b2b-copy
by daniel-ospina
Generates carousel copy and creative direction from a creative brief. Invoked by carousel-b2b-strategy. Runs internal review loop (tone, brand, structure) before presenting script to user.
config-validation
by daniel-ospina
"Thin wrapper that runs relevant check scripts based on changed files. Maps file types to validation scripts (migrations → check-migration-*, skills → check-skill-lint, etc.). Invoked by test-routing when domain=config."
content-reviewer-depth
by daniel-ospina
Use when reviewing editorial, guide, or deal content for local depth, unique insights, and audience consistency. Checks that content goes beyond generic travel blog coverage, incorporates unique research findings, and speaks consistently to the right audience (residents for ES, expats/tourists for EN). Returns structured issues and specific fixes. Part of content pipeline v2 post-writing review — run in parallel with other reviewers.
content-reviewer-breadth
by daniel-ospina
Use when reviewing editorial, guide, or deal content to ensure nothing essential from the research brief was dropped during writing. Checks neighborhood coverage, cuisine/experience types, price tiers, and partner completeness. Returns structured issues and specific fixes. Part of content pipeline v2 post-writing review — run in parallel with other reviewers.
content-fact-checker-writing
by daniel-ospina
Use after the text reviewer cycle is clean, before final delivery. Cross-checks all factual claims in the written content against the verified research brief. No Perplexity calls — only cross-references against verified research. Returns issues in standard P0/P1/P2 reviewer format.
content-verification
by daniel-ospina
"Content-domain verification gate. Dispatches the content pipeline v2 reviewers — breadth + depth in parallel (post-writing), then fact-checker (writing) after the reviewer cycle is clean — for editorial, guide, deal, and refresh content. Invoked by test-routing when domain=content; referenced by research-verification and issue-creation as the Content verification trigger. Returns structured P0/P1/P2 issues + verdict."
content-staleness-scanner
by daniel-ospina
Use when scanning for stale editorial content that needs refreshing, when asked to "check for stale content", "find pages that need refreshing", "run staleness scan", or "what content needs a refresh". Queries the DB, outputs a prioritized refresh queue, and optionally triggers content-refresh for each item.
debug-workflow
by daniel-ospina
Pipeline-gated debugging workflow. Classifies bug complexity, routes through issue-workflow to task-workflow (micro), task-workflow-standard or project-workflow (standard/complex), then applies systematic root-cause methodology. Use when encountering any bug, test failure, or unexpected behavior.