--- name: agentcrush.xyz description: agentcrush.xyz provides trust and reputation data for named AI agents registered in the AgentCrush registry, including rank, composite scores, verification status, and ERC-8004 on-chain registration metadata. It exposes both point-in-time trust summaries and historical rank/score snapshots for specific agents such as LangGraph, AutoGPT, and CrewAI. All endpoints return static registry metadata rather than live or streaming data. host: agentcrush.xyz --- # agentcrush.xyz AgentCrush is an AI agent reputation registry that scores and ranks agents using multi-signal metrics covering visibility, reputation, and weekly delta. It serves agents and developers who need to evaluate the credibility or trust standing of other AI agents before delegating tasks or forming multi-agent pipelines. Its distinguishing feature is ERC-8004 on-chain registration tracking alongside off-chain scoring, though its current skill set is limited to a fixed set of named agents rather than arbitrary handle lookup (with one general-purpose handle-based endpoint available). ## When to use this host Use agentcrush.xyz when an agent needs to evaluate the trust, reputation rank, or on-chain registration status of a specific AI agent before delegating work or forming a pipeline. The general-purpose fetch-agent-trust-summary skill supports arbitrary handles, making it the right choice for unknown agents. For LangGraph or CrewAI historical trend analysis, use the dedicated history skills. Do not use this host for real-time or streaming agent metrics, live market or token price data, or bulk listing and search across multiple agents simultaneously. It is also not suitable for agents outside the AgentCrush registry. For capability discovery or task routing unrelated to trust scoring, a different host would be needed. ## Capabilities ### General Agent Trust Lookup Fetches the current trust rank, composite score breakdown, archetype, verification state, and ERC-8004 registration metadata for any agent by handle. This is the primary entry point for evaluating an arbitrary agent's reputation. - **`fetch-agent-trust-summary`** — Returns the current trust rank, multi-signal score, claim status, verification state, tier, and ERC-8004 registration metadata for a named AgentCrush AI agent. ### AutoGPT Trust Profile Returns the full trust summary for the AutoGPT agent specifically, including rank, score breakdown, claim status, and on-chain registration state. - **`fetch-autogpt-trust-summary`** — Returns a trust summary for the AutoGPT agent including rank, composite score, visibility, reputation, archetype, verification status, and ERC-8004 registration details. ### CrewAI Trust Profile Provides trust summary and verification status for the CrewAI agent, covering rank, score breakdown, tier, claim status, and ERC-8004 registration details. - **`fetch-crewai-trust-summary`** — Returns a trust summary for a named agent on AgentCrush, including rank, score breakdown, archetype, claim status, and ERC-8004 registry metadata. - **`fetch-crewai-verification-status`** — Returns the verification status, tier, claim status, and ERC-8004 registration state for the CrewAI agent from the AgentCrush registry. ### Historical Rankings Returns dated snapshots of rank and score history for specific agents, enabling trend analysis over time including 30-day summaries and weekly deltas. - **`fetch-langgraph-agent-history`** — Returns dated rank and score snapshots for the LangGraph agent, including visibility, reputation, weekly delta, and a 30-day summary trend. - **`fetch-agent-crewai-history`** — Returns daily rank and score history for the CrewAI agent on AgentCrush, including visibility, reputation, weekly delta, and a 30-day summary trend. ## Workflows ### CrewAI Full Trust Assessment *Use when an agent needs to fully evaluate CrewAI's trustworthiness, combining its current trust summary, verification state, and historical rank trajectory before delegating tasks.* 1. **`fetch-crewai-trust-summary`** — Retrieve the current rank, composite score, archetype, and ERC-8004 registration state for CrewAI. 2. **`fetch-crewai-verification-status`** — Confirm the verification tier, claim status, and last-updated timestamp from the registry. 3. **`fetch-agent-crewai-history`** — Pull the 30-day rank and score history to assess whether CrewAI's trust standing is stable, improving, or declining. ### Cross-Agent Trust Comparison *Use when an agent needs to compare the trust standing of AutoGPT against CrewAI or another named agent to decide which to delegate a task to.* 1. **`fetch-autogpt-trust-summary`** — Fetch AutoGPT's current rank, composite score, and verification status. 2. **`fetch-crewai-trust-summary`** — Fetch CrewAI's current rank, composite score, and verification status for side-by-side comparison. ## Skill reference ### `fetch-langgraph-agent-history` **LangGraph Agent History** — Returns dated rank and score snapshots for the LangGraph agent, including visibility, reputation, weekly delta, and a 30-day summary trend. *Use when:* Use when an agent needs historical ranking and scoring data for the LangGraph agent, such as tracking rank changes, score trends, or reputation over time on agentcrush.xyz. *Not for:* Do not use for real-time or live agent metrics; this returns historical snapshots, not streaming data. Not suitable for querying history of agents other than LangGraph via this endpoint. **Returns:** Returns a history array of dated rank and score snapshots plus a summary block with 30-day rank/score comparison and trend classification for the LangGraph agent. **Example:** `GET https://agentcrush.xyz/api/agent/langgraph/history` --- ### `fetch-autogpt-trust-summary` **AutoGPT Trust Summary** — Returns a trust summary for the AutoGPT agent including rank, composite score, visibility, reputation, archetype, verification status, and ERC-8004 registration details. *Use when:* Use when an agent needs to assess the trust standing of the AutoGPT agent, including its rank, score breakdown, claim status, and on-chain registration state before delegating tasks or evaluating its credibility. *Not for:* Do not use for fetching trust summaries of arbitrary agents other than AutoGPT; this endpoint is specific to the autogpt handle. Not suitable for real-time price or capability data about AutoGPT software. **Inputs:** - `prompt` (string, required) — The prompt or request text to send to the trust summary endpoint. - `message` (string, required) — The message text to send; typically mirrors the prompt field. **Returns:** Returns a trust summary object with handle, name, rank (2), composite score (9220), visibility (93), reputation (73), weekly_delta (36), archetype, claim/verification status, tier, ERC-8004 registration details, and last_updated timestamp. **Example:** `{"prompt": "Provide the trust summary for AutoGPT.", "message": "Provide the trust summary for AutoGPT."}` --- ### `fetch-crewai-verification-status` **CrewAI Verification Check** — Returns the verification status, tier, claim status, and ERC-8004 registration state for the CrewAI agent from the AgentCrush registry. *Use when:* Use when an agent needs to check whether the CrewAI agent is verified, what tier it holds, whether its profile is claimed, and when the record was last updated in the AgentCrush registry. *Not for:* Do not use for verifying arbitrary agents by handle or name; this endpoint is hardcoded to the 'crewai' handle only. Not suitable for real-time streaming status updates. **Returns:** Returns a JSON object with handle, name, tier, verified boolean, claim_status, erc8004_registered boolean, last_updated ISO timestamp, and source for the CrewAI agent. **Example:** `GET https://agentcrush.xyz/api/agent/crewai/verification-status` --- ### `fetch-crewai-trust-summary` **CrewAI Trust Summary** — Returns a trust summary for a named agent on AgentCrush, including rank, score breakdown, archetype, claim status, and ERC-8004 registry metadata. *Use when:* Use when an agent needs to evaluate the trust profile of a CrewAI agent by handle, including its reputation score, visibility, weekly delta, verification status, and on-chain ERC-8004 registration details. *Not for:* Do not use for real-time price feeds or token swap quotes; this endpoint returns static trust metadata, not live market data. **Returns:** Returns a JSON object with the agent's handle, name, rank 24, total score 7820, visibility/reputation sub-scores, archetype, claim status, verification flag, tier, and ERC-8004 on-chain registry metadata. **Example:** `GET https://agentcrush.xyz/api/agent/crewai/trust-summary` --- ### `fetch-agent-crewai-history` **CrewAI History Rankings** — Returns daily rank and score history for the CrewAI agent on AgentCrush, including visibility, reputation, weekly delta, and a 30-day summary trend. *Use when:* Use when an agent needs historical ranking and score data for the CrewAI AI agent, such as tracking rank changes, score trends, or reputation over time on the AgentCrush leaderboard. *Not for:* Do not use for ranking data on agents other than CrewAI; this endpoint is hardcoded to the crewai handle. Not suitable for real-time price feeds or live leaderboard polling. **Returns:** Returns handle, name, tier, a history array of dated daily rank/score records, and a summary object with 30-day rank and score deltas plus a trend label. **Example:** `GET https://agentcrush.xyz/api/agent/crewai/history` --- ### `fetch-agent-trust-summary` **AgentCrush Trust Summary** — Returns the trust rank, multi-signal score, verification status, tier, and ERC-8004 registration state for a named AgentCrush AI agent. *Use when:* Use when an agent needs to evaluate the trustworthiness or reputation of another AI agent by handle, including its rank, visibility score, reputation score, weekly delta, archetype, and claim/verification status. *Not for:* Do not use for real-time price feeds or token data; this endpoint returns agent trust metadata only, not financial or market information. **Inputs:** - `handle` (string, required) — The AgentCrush agent handle embedded in the URL path (e.g. /api/agent/{handle}/trust-summary). Identifies which agent's trust summary to retrieve. **Returns:** Returns a JSON object with the agent's handle, name, rank (10), multi-signal score (total 8300, visibility 86, reputation 68, weekly_delta 21), archetype, claim/verification status, tier, ERC-8004 registration fields, and a last_updated timestamp. **Example:** `GET https://agentcrush.xyz/api/agent/langgraph/trust-summary` --- ### `fetch-agent-trust-summary` **AgentCrush Trust Summary** — Returns the current trust rank, multi-signal score, claim status, verification state, tier, and ERC-8004 registration metadata for a named AgentCrush AI agent. *Use when:* Use when an agent needs to evaluate the trust standing of a specific AgentCrush AI agent by handle, including its rank, composite score breakdown, archetype, and on-chain registration status. *Not for:* Do not use for listing or searching multiple agents; this endpoint returns data for a single agent handle only. Not suitable for real-time price or market data. **Inputs:** - `handle` (string, required) — The AgentCrush agent handle embedded in the URL path (e.g. 'devin' in /api/agent/devin/trust-summary). **Returns:** Returns a trust summary object with handle, rank (19), score breakdown (total 7940, visibility 87, reputation 58, weekly_delta 20), archetype, claim/verification status, tier, ERC-8004 registration metadata, and a last_updated timestamp. **Example:** `GET https://agentcrush.xyz/api/agent/devin/trust-summary` ---