Agent
Compare options by speed, risk, evidence and reversibility.
Agent
Compare options by speed, risk, evidence and reversibility.
Agent
Create testable outreach angles tied to pains and proof.
Agent
Collect account candidates with source caveats and no fake data.
Agent
Define tight ICP and exclusion criteria.
Agent
Quantify customer impact, affected systems, duration and residual risk.
Agent
Write the RCA, immediate mitigation, verification steps and owner-bound follow-ups.
Agent
Build an ordered timeline from incident notes, alerts, deploys and operator actions.
Agent
Check test coverage, monitoring, support path, migration safety and rollback criteria.
Agent
Write the final launch verdict, rollout steps and remaining blocked items.
Agent
Extract launch scope, target users, dependencies, user-visible copy and excluded work.
Agent
Extract the decision, target user, claims and requested outcome.
Agent
Give the strongest next product move with scope boundaries.
Agent
Extract exact claims that need verification.
Agent
Find code, docs and command evidence; do not trust summaries.
Agent
Give a direct verdict with evidence and caveats.
Agent
Map realistic misuse paths from the declared capabilities.
Agent
Write a permission verdict and least-privilege mitigation plan.
Agent
Read manifest permissions, declared tools, secrets, network rules and approval gates.
Agent
Classify urgency, ownership, money impact and required evidence.
Agent
Draft a concise support response without unsupported promises.
Agent
Preserve exact customer statements and dates.
Agent
You run the harness workflow.
MCP server Claude CodeCodexCursor +2
Find, inspect and install reviewed reusable AI-agent capabilities from superskill.sh. Remote server at superskill.sh.