Skip to main content
Glama
Sign In
agentflight

agentflight

by codesphere-u
OverviewSchemaSourceRelated ServersDiscussionsAdmin
Python
Local

Server Configuration

Describes the environment variables required to run the server.

NameRequiredDescriptionDefault

No arguments

Instructions

Guidance the server publishes about itself, which clients place ahead of the tool catalog so the model reads it before choosing anything.

This server publishes no instructions, or was last inspected before Glama recorded them.

Capabilities

Features and capabilities supported by this server

Protocol revision2025-11-25

CapabilityDetails
tools
{
  "listChanged": false
}
prompts
{
  "listChanged": false
}
resources
{
  "subscribe": false,
  "listChanged": false
}

Tools

Functions exposed to the LLM to take actions

NameDescription
agentflight_diff_cassettesA

Compare a baseline golden AgentFlight cassette against a candidate cassette to detect behavioral regressions.

Purpose

Performs semantic trajectory diffing across model, prompt, or tool definition upgrades. Evaluates tool sequence divergence using Levenshtein distance and Longest Common Subsequence (LCS), identifies newly introduced tool failures, flags recursive loop regressions, and calculates token spend and USD cost deltas.

When to use

  • After upgrading LLM foundation models, prompt templates, or system instructions to verify agent behavior parity.

  • In CI/CD pipelines to block regressions when agent trajectory similarity drops below threshold or costs spike.

When NOT to use

  • Do NOT use to inspect single cassette integrity; use agentflight_verify_cassette instead.

  • Do NOT use to convert cassettes to telemetry traces; use agentflight_export_otel instead.

Behavior & Side Effects

  • Read-only: Reads both cassette files from disk without altering them.

  • Deterministic & Idempotent: Repeated invocations with identical cassettes yield exact identical diff metrics.

  • Returns structured diff metrics including similarity score (0.0 to 1.0), regression alerts, and cost deltas.

agentflight_export_otelA

Export recorded AgentFlight cassette spans into W3C TraceContext and OpenTelemetry JSON traces.

Purpose

Converts recorded agent execution spans into OpenTelemetry-compliant trace spans with standard gen_ai semantic conventions, enabling critical-path latency waterfall visualization and distributed tracing in systems like StackPulse, Jaeger, or Datadog.

When to use

  • Exporting agent trajectories into APM or observability platforms for SLO tracking and latency bottleneck analysis.

  • Visualizing parent-child relationships and parallel tool execution spans.

When NOT to use

  • Do NOT use to verify hash chain cryptographic integrity; use agentflight_verify_cassette instead.

  • Do NOT use to compare two trajectories for regressions; use agentflight_diff_cassettes instead.

Behavior & Side Effects

  • If output_path is provided, writes the formatted OpenTelemetry JSON to that file; otherwise returns spans in-memory.

  • Idempotent: Successive calls with identical arguments produce identical span outputs.

agentflight_lint_toolsA

Statically inspect MCP tool schema definitions against 10 ergonomics rules (TS001-TS010) and calculate an Agent Ergonomics Score.

Purpose

Evaluates Model Context Protocol tool schemas for anti-patterns that confuse LLM tool-calling (tautological descriptions, missing required arrays, ambiguous parameter types, unconstrained inputs, and lack of usage boundaries). Produces an ergonomics score (0-100) and actionable remediation steps.

When to use

  • Auditing and validating MCP tool definitions during development or CI/CD pre-commit hooks.

  • Benchmarking tool schema clarity and LLM tool selection readiness before deployment.

When NOT to use

  • Do NOT use to evaluate runtime agent trajectory recordings; use agentflight_diff_cassettes or agentflight_verify_cassette instead.

Behavior & Side Effects

  • Read-only: Operates purely in-memory; does not mutate any schemas or files.

  • Deterministic: Analyzes schemas through static AST and JSON Schema inspection with zero external network calls.

agentflight_verify_cassetteA

Cryptographically verify the SHA-256 hash chain and structural integrity of an AgentFlight cassette.

Purpose

Validates that a recorded agent execution cassette (.flight.json or .flight.jsonl) has not been tampered with, truncated, or corrupted. Verifies that every span links to the previous span digest via canonical JSON SHA-256 hashing.

When to use

  • Auditing compliance and tamper-proof provenance of recorded agent trajectories.

  • Pre-flight verification before running deterministic replays with agentflight.replayer.

When NOT to use

  • Do NOT use to compare two cassettes against each other; use agentflight_diff_cassettes instead.

  • Do NOT use to lint MCP tool definitions; use agentflight_lint_tools instead.

Behavior & Side Effects

  • Read-only: Reads cassette from disk and recalculates the SHA-256 chain in-memory without modifications.

  • Deterministic: Always outputs exact boolean verification status, step count, and cryptographic head hash.

Prompts

Interactive templates invoked by user choice

NameDescription
lint-tool-definitionsReview and harden MCP tool definitions against the 10 AgentFlight tool-smell rules (TS001-TS010) to maximize LLM tool-calling accuracy.
review-agent-trajectoryArchitectural post-mortem prompt to analyze an AI agent trajectory for repetitive loops, runaway token spend, or tool errors.

Resources

Contextual data attached and managed by the client

NameDescription
MCP Tool-Smell Linter Rule Catalog (TS001-TS010)Comprehensive specification of the 10 static ergonomics rules for MCP tool definitions, including severity, antipattern signatures, and suggested fixes.
AgentFlight Performance Benchmarks & Subsystem SpecsDeterministic trajectory replay latency (<0.02ms), tool-smell linter throughput (1,676 tools/sec), and zero-overhead circuit-breaker benchmarks.
Deploy Server
Claude Code, inside the Godot 4 editorFlockbay puts an AI agent in the real Godot 4 editor: it writes GDScript, runs your game and fixes what breaks. Free, or bring your own Claude Code or Codex.flockbay.comAd

TDQS

A4.6/5.0

Scored across 4 tools

Disambiguation5/5

Each tool has a clearly distinct action and output: verify (crypto integrity), diff (behavioral regression), export (OTel traces), and lint (schema ergonomics). Descriptions even include explicit "When NOT to use" cross-references pointing to the correct sibling tool, eliminating overlap. An agent can select unambiguously from names alone.

Naming Consistency5/5

All four tools follow an identical `agentflight_verb_noun` snake_case pattern: diff_cassettes, export_otel, lint_tools, verify_cassette. The shared server prefix plus a clear action verb keeps the set predictable and readable.

Tool Count5/5

Four tools is a tight, well-scoped surface for a cassette verification/observability toolkit, and each tool maps to a distinct lifecycle stage rather than duplicating functionality. Nothing appears padded or redundant.

Completeness4/5

The analysis/verification side of the domain is well covered (verify, diff, export, lint), but there is no tool to record, list, or replay cassettes — replay is delegated to an external `agentflight.replayer` CLI. This is a workable but noticeable gap for an agent trying to run a full record-verify-diff-replay loop from MCP alone.

Maintenance

ActivityMaintained
ResponsivenessNo recent community issues

Resources