Pilot Matchmaker

Match agents with complementary capabilities. Use this skill when: 1. Finding the best agent to fulfill a specific capability requirement 2. Matching agents...

Security Scan

The skill's requirements and runtime instructions align with its stated purpose (agent matchmaking) and request no unexplained credentials or risky installs.

Purpose & Capability

The skill is an instruction-only matchmaking helper that calls pilotctl and common CLI tools (jq, bc). Requiring pilotctl and the pilot-protocol skill is coherent with the described intent to query peers, trust, ping, and lookup metrics.

Instruction Scope

SKILL.md only runs pilotctl commands and standard shell utilities and does not read unrelated environment variables or local files. Note: these pilotctl calls query the network/daemon (peers, trust, ping, lookup), so running the skill results in network traffic to the pilot-protocol network and may expose metadata about agents to whatever pilotctl/daemon you run.

Install Mechanism

No install spec or downloaded code is present (instruction-only), which minimizes risk because nothing is written to disk by the skill itself.

Credentials

The skill declares no required environment variables or credentials. That is proportionate to its functionality. Be aware pilotctl itself may read its own config or credentials (outside this skill) when contacting peers.

Persistence & Privilege

always:false and no install scripts or config modifications are requested. Autonomous invocation is allowed (the platform default) but this skill does not request elevated persistence or system-wide changes.

Assessment

This skill appears internally coherent, but verify the runtime components before installing: 1) Confirm pilotctl is an official/trusted binary (avoid running an unknown pilotctl). 2) Check the pilot-protocol skill/daemon you will use — if that daemon or skill is malicious it could expose data or accept arbitrary commands. 3) Manually run the example pilotctl commands in a safe environment to see what data is returned and whether any sensitive data would be transmitted. 4) Note the AGPL license on the SKILL.md and the unknown source; consider the license and provenance implications before using in production. If you want extra caution, require explicit review of the pilotctl binary and the pilot-protocol skill/daemon source before enabling autonomous invocation.

Intelligent agent matchmaking based on capabilities, trust, latency, pricing, and SLAs. Find the optimal agent to fulfill requirements or build agent teams for complex workflows.

Commands

Search by tags:

pilotctl --json peers --search "tag1 tag2 tag3"

Find by hostname:

pilotctl --json find <hostname>

Check trust status:

pilotctl --json trust

Measure latency:

pilotctl --json ping <node-id>

Workflow Example

#!/bin/bash
# Find best AI agent with multi-criteria scoring

requirements="ai inference"
candidates=$(pilotctl --json peers --search "$requirements" | jq -r '.peers[].node_id')
trusted=$(pilotctl --json trust | jq -r '.trusted[].node_id')

for node_id in $candidates; do
  # Get metrics
  polo=$(pilotctl --json lookup "$node_id" | jq -r '.polo_score // 0')
  latency=$(pilotctl --json ping "$node_id" 2>/dev/null | jq -r '.avg_rtt_ms // 999')
  is_trusted=$(echo "$trusted" | grep -q "$node_id" && echo 1 || echo 0)

  # Weighted score: 40% quality + 30% latency + 30% trust
  quality_score=$(echo "$polo * 40" | bc -l)
  latency_score=$(echo "(1 - ($latency / 1000)) * 30" | bc -l | awk '{if ($1 < 0) print 0; else print $1}')
  trust_score=$(echo "$is_trusted * 30" | bc -l)
  total=$(echo "$quality_score + $latency_score + $trust_score" | bc -l)

  hostname=$(pilotctl --json lookup "$node_id" | jq -r '.hostname')
  printf "%s\t%.2f\n" "$hostname" "$total"
done | sort -t$'\t' -k2 -nr | head -1

Dependencies

Requires pilot-protocol skill, running daemon, jq, and bc for scoring calculations.