Public agent skills by SwiftAdviser.
These skills are written as SKILL.md contracts for tool-using agents such as Codex, OpenAI agents, Claude Code, Cursor, Cline, Amp, and other runtimes that support agent skills. They are distributed through skills.sh, not a Claude-only marketplace.
Private app source code, deployment configuration, environment files, API keys, tokens, and operational secrets do not belong in this repository.
- Installation
- Available Skills
- bestchange - BestChange exchanger discovery
- monobank - Monobank balance checks
- uah-mono-payouts - USDT BEP20 to UAH payout flow
- ai-hypergrowth-gtm - AI sales-led hypergrowth GTM
- high-quality-content-writer - Anti-slop content writing and scoring
- mindwtr - Mindwtr Cloud and recurring task CLI
- Repository Structure
- License
List the catalog first because each install command targets one named skill globally. Every install uses the same SKILL.md contract from this repository, so supported agent runtimes read the same operating instructions.
List available skills:
npx skills add SwiftAdviser/public-skills --listInstall one skill globally:
npx skills add SwiftAdviser/public-skills --skill bestchange -g -y
npx skills add SwiftAdviser/public-skills --skill monobank -g -y
npx skills add SwiftAdviser/public-skills --skill uah-mono-payouts -g -y
npx skills add SwiftAdviser/public-skills --skill ai-hypergrowth-gtm -g -y
npx skills add SwiftAdviser/public-skills --skill high-quality-content-writer -g -y
npx skills add SwiftAdviser/public-skills --skill mindwtr -g -yFind trusted BestChange exchangers for crypto-to-fiat and e-currency exchange requests through the hosted BestChange MCP endpoint.
- Resolves exact BestChange currency codes before quoting
- Returns ranked exchanger options for a resolved pair
- Explains empty results with diagnostics instead of guessing
- Keeps live rates behind the MCP tool instead of stale README examples
MCP endpoint:
https://bestchange-mcp.krutovoy.me/mcp
Triggers:
- "bestchange options"
- "find exchangers for USDT to UAH"
- "crypto to fiat exchanger ranking"
- "20 usdc base to revolut eur"
Answer Monobank balance questions by calling the official Monobank API directly with a user-supplied per-request token.
- Calls the official Monobank client-info endpoint directly
- Does not route real balances through a hosted MCP proxy
- Masks account identifiers in user-facing output
- Handles Monobank rate limits and token errors
API endpoint:
https://api.monobank.ua/personal/client-info
Triggers:
- "скільки грошей у мене на монобанку?"
- "баланс mono"
- "monobank accounts"
- "monobank balance"
Convert USDT BEP20 into UAH payout instructions through a verified exchange flow with quote, approval, expiry, AML screening, and payment monitoring.
- Searches contacts from agent-owned local context
- Quotes USDT BEP20 to UAH payout routes
- Creates orders only after explicit user approval
- Renders operational payment instructions
- Polls order status and fails closed on unsafe states
MCP endpoint:
https://mcp-wallet.mandate.md/mcp
Triggers:
- "выведи 500 USDT BEP20 на mono UAH"
- "convert USDT BEP20 to UAH payout"
- "create UAH payout order"
- "check payout status"
Apply AI sales-led hypergrowth patterns to B2B startup GTM: wedge selection, founder-led sales, design partners, paid pilots, labor-cost pricing, enterprise expansion, and agent-infrastructure platform-risk analysis.
- Turns hypergrowth research into concrete GTM decisions
- Includes Jina Reader markdown references for the public research corpus
- Maps startup situations to company cases, growth laws, and sales laws
- Produces founder-owned next actions, pilot offers, and pricing anchors
- Stress-tests agent-infrastructure ideas against platform/distribution risk
Research source:
https://sevaustinov.me/hypergrowth-research/
Triggers:
- "evaluate our GTM wedge"
- "design a paid pilot"
- "how should founders sell first enterprise deals?"
- "price this AI product against labor cost"
- "does this agent infrastructure wedge survive platform risk?"
Write, rewrite, audit, and improve prose through deterministic anti-slop gates. Use it for content cleanup, link or text grading, AI-writing trope removal, article/copy review, and closed-loop rewrites until the draft is specific, fact-dense, and publishable.
- Scores pasted text, local files, and URLs for AI-writing trope density
- Runs a bundled quality gate with specificity, causality, voice, and slop-risk criteria
- Creates retry artifacts for evaluator and writer subagents
- Preserves meaning and facts while removing generic claims, filler, and unsupported polish
- Includes public references for the scoring rubric and isolated writer/evaluator flows
Primary scripts:
scripts/slop_score.py
scripts/quality_gate.py
scripts/content_quality_loop.py
Triggers:
- "check this text for AI slop"
- "rewrite this until it passes quality gates"
- "score this article"
- "remove AI-writing tropes"
- "make this copy specific and publishable"
Manage tasks in any authenticated Mindwtr Cloud instance from Codex, OpenClaw, or a shell, including recurring schedules that are not exposed by the current Mindwtr MCP schemas.
- Uses the Mindwtr Cloud REST API with bearer-token or mode-0600 token-file authentication
- Creates strict calendar or fluid after-completion recurrence
- Supports list, search, create, update, complete, archive, and guarded soft-delete operations
- Includes dry-run recurrence payloads, routing fixtures, and deterministic CLI logic
- Keeps server URLs and tokens out of the public skill
Primary CLI:
scripts/mindwtr.mjs
Triggers:
- "create a Mindwtr task"
- "add a recurring task every Monday"
- "list my Mindwtr tasks"
- "создай задачу в Mindwtr"
- "добавь повторяющуюся задачу"
public-skills/
README.md
skills/
bestchange/
SKILL.md
monobank/
SKILL.md
uah-mono-payouts/
SKILL.md
ai-hypergrowth-gtm/
SKILL.md
references/
high-quality-content-writer/
SKILL.md
references/
scripts/
mindwtr/
SKILL.md
references/
scripts/
Each skill folder is intended to be installable on its own. SKILL.md is the full agent contract; reference files are loaded only when the agent needs deeper context.
MIT