riptide

Assess a Solana program with Riptide — run deterministic, guided LiteSVM simulations against the real on-chain program to produce a risk assessment with reproducible evidence. Use when the user says "assess my protocol", "is my protocol safe", "run Riptide on this", "give me a risk assessment", or "riptide-assess", or points you at an Anchor/Solana lending, AMM, perps, liquid-staking, or stablecoin repo and wants one agent-led flow from program source to an assessment report. Detects the protocol family, scopes what the guided sim must handle (typed args, Pyth/Switchboard oracle-account bytes, liquidator/keeper actors, multi-instruction sequences), authors and runs the sim, then returns assessment evidence and exact rerun commands.

One front door for Riptide protocol assessments. Use when the user says "riptide-assess", "assess my protocol", "is my protocol safe", "run Riptide on this", "give me a risk assessment", or wants one agent-led flow from a Solana program repo to an assessment report. This skill detects the protocol family, scopes the risk, asks up to three scoped questions, delegates the guided-sim setup to riptide-config, runs the guided-sim evidence flow (sim generate, sim run, sim surface), and returns assessment evidence with exact rerun commands. For deeper guided-sim authoring without a final assessment, use riptide-config instead.

One front door for Riptide protocol assessments. Use when the user says "riptide-assess", "assess my protocol", "is my protocol safe", "run Riptide on this", "give me a risk assessment", or wants one agent-led flow from a Solana program repo to an assessment report. This skill detects the protocol family, scopes the risk, asks up to three scoped questions, delegates the guided-sim setup to riptide-config, runs the guided-sim evidence flow (sim generate, sim run, sim surface), and returns assessment evidence with exact rerun commands. For deeper guided-sim authoring without a final assessment, use riptide-config instead.

Configure a Solana program repo for Riptide end to end after the thin default `riptide init` bootstrap. Use when the user says "riptide-config", "configure Riptide", "make this repo run in Riptide", "finish my adapter/sim", or has a `.riptide/` scaffold that needs a working adapter, a generated guided-sim crate, filled setup seams, a parameter sweep, validation, and a final readiness verdict. This merged skill owns the adapter, guided-sim setup, sweep, repair, and assessment readiness loop.

Synthesize a case-study-voice narrative markdown report for a Riptide simulation run using the current agent session. Use when the user says "narrative report", "riptide narrative", "turn this run into prose", "write up this simulation", or is standing in a directory with a `simulation-result.json` + `report.md` pair and wants a richer, causal retelling that cites specific ticks, events, and invariant firings. The skill reads both artifacts, drafts the narrative in-session, and writes `report-narrative.md` alongside the existing `report.md`. Zero setup, no API keys, no endpoint config.

Synthesize a case-study-voice narrative markdown report for a Riptide simulation run using the current agent session. Use when the user says "narrative report", "riptide narrative", "turn this run into prose", "write up this simulation", or is standing in a directory with a `simulation-result.json` + `report.md` pair and wants a richer, causal retelling that cites specific ticks, events, and invariant firings. The skill reads both artifacts, drafts the narrative in-session, and writes `report-narrative.md` alongside the existing `report.md`. Zero setup, no API keys, no endpoint config.