← Back to documentation

BUY-2700_ai_tool_directory_submission_sprint_board_2026-04-16

BUY-2700: AI Tool Directory Submission Sprint Board and Retry Plan

Owner: Reach Prepared: 2026-04-16 Sprint window: 2026-04-17 to 2026-04-19

Objective

Turn the current AI-tool directory backlog into an executable board that:

  • keeps only live, high-fit destinations
  • captures the submission mechanics or blocker for each destination
  • separates agent-executable work from manual or board-dependent work
  • routes follow-up work into the right growth lane without more rediscovery

Current State

Completed and should not be retried in this sprint:

  • Glama.ai
  • Smithery.ai
  • Anthropic MCP Registry

Live validation completed on 2026-04-16:

  • mcp.directors no longer resolves and should be removed from active plans
  • superagent.sh currently resolves to an AI security company, not a public registry destination for this workflow
  • agentops.ai is live, but the public site does not expose a clear self-serve listing flow; its footer points to the Agency marketplace at marketplace.agen.cy
  • OpenInterpreter/open-interpreter is an active GitHub project, but the repo does not expose a registry submission path; treat this as ecosystem integration or outreach, not a directory submission
  • punkpeye/awesome-mcp-servers is active and has a clear contribution path via GitHub PR; the README states the web directory is synced with Glama

Sprint Board

PriorityDestinationCurrent statusSubmission pathRequirementsExecution modeNext actionOwner lane
1punkpeye/awesome-mcp-serversLive, high-fit, executableGitHub PR editing README.md under the right categoryRepo URL, one-line description, category placement, contribution-format complianceAgent-executablePrepare PR packet using existing MCP metadata and submit firstDistribution
2Glama sync checkAlready listed; no new submission neededVerify the existing glama.json metadata still matches the current product storyMetadata review onlyAgent-executableUse as canonical source for all other directory copyDistribution
3Smithery sync checkAlready listed; no new submission neededVerify smithery.yaml metadata still matches the current product storyMetadata review onlyAgent-executableUse as canonical source for install steps and env var wordingDistribution
4Agency marketplace (marketplace.agen.cy)Live but unclear listing mechanicsNo visible self-serve submit route found from public pagesHuman-visible browser and/or direct contact pathManualRoute to manual-outreach lane with a ready listing packet instead of attempting blind automationGrowth ops
5AgentOps / Agency contactLive company site, but not a confirmed directoryPublic contact path onlyHuman follow-up, likely intro note or partnership-style askManualTreat as outreach, not directory automationGrowth ops
6OpenInterpreter ecosystemActive project, but not a registryGitHub/community integration path still needs a specific askDecide whether the follow-up is docs, example, issue, or integration PRMixedReframe from "registry submission" to "ecosystem integration brief"DevRel
7mcp.directorsDeadNoneDNS failureRemoveDelete from active queue and execution logsArchive
8superagent.shStale for this use caseNone for directory listingDomain now represents a different product categoryRemoveDelete from active queue and execution logsArchive

Three-Day Plan

2026-04-17

Primary goal: execute the one clean GitHub-native listing path and lock canonical metadata.

  1. Normalize canonical listing copy from:
    • glama.json
    • smithery.yaml
    • claude-mcp-registry-submission.md
  2. Prepare the awesome-mcp-servers entry:
    • category recommendation: E-Commerce
    • short description: BuyWhere as an agent-native product search and price-comparison MCP server
    • repository URL: https://github.com/buywhere/buywhere-api
  3. Submit or stage the PR packet.
  4. Mark mcp.directors and superagent.sh as invalid in all active sprint references.

2026-04-18

Primary goal: hand off manual-only targets with zero rediscovery required.

  1. Package the Agency marketplace / AgentOps listing brief:
    • product name
    • one-line description
    • long description
    • install steps
    • MCP use case
    • links and UTM targets
  2. Route that packet to the manual growth lane for browser-visible submission or contact outreach.
  3. Reframe OpenInterpreter from registry submission to one of:
    • ecosystem issue
    • example integration
    • docs mention
    • partner outreach

2026-04-19

Primary goal: clean the backlog and prevent stale retries.

  1. Update the standing target lists so stale entries are not re-queued.
  2. Leave the next operator with:
    • one completed or in-flight GitHub-native submission
    • one manual packet for Agency marketplace / AgentOps
    • one reframed follow-up brief for OpenInterpreter

Retry Rules

Retry only when the destination still has a confirmed route.

Retry now:

  • awesome-mcp-servers if the first PR needs formatting or category changes
  • manual Agency marketplace / AgentOps outreach once a human-visible path is available

Do not retry until a new route is discovered:

  • mcp.directors
  • superagent.sh

Do not classify as "directory submission" without a new proof point:

  • OpenInterpreter

Canonical Listing Packet

Use this copy across GitHub-native and manual destinations unless the target forces shorter fields.

Product name:

  • BuyWhere

Short description:

  • Agent-native product catalog API for shopping agents. Search products, compare prices, and retrieve structured commerce data through one MCP-ready API.

Long description:

  • BuyWhere is the retrieval layer for AI commerce workflows. It gives developers and agent builders structured product search, price comparison, deal discovery, and category browsing instead of retailer-by-retailer scraping. The MCP server and API are designed for agent workflows, with installable tools for search, product lookup, comparison, deals, and category navigation.

Repository:

  • https://github.com/buywhere/buywhere-api

Homepage:

  • https://buywhere.ai

Docs:

  • https://api.buywhere.ai/docs

Primary MCP command:

python mcp_server.py

Required environment variables:

  • BUYWHERE_API_KEY
  • BUYWHERE_API_URL defaulting to https://api.buywhere.ai

Lane Routing

Distribution lane:

  • awesome-mcp-servers PR
  • metadata parity check for Glama and Smithery

Growth ops lane:

  • Agency marketplace submission attempt
  • AgentOps / Agency contact follow-up

DevRel lane:

  • OpenInterpreter ecosystem integration brief

Archive lane:

  • remove mcp.directors
  • remove superagent.sh
  • update legacy docs that still describe both as active targets

Source Links

Validated on 2026-04-16: