Companion to #16. Tracks Glama.ai coverage for the *-mcp fleet. Surfaced by stars-watcher's daily digest — currently reporting 5 of 45 *-mcp repos indexed by Glama.
How Glama actually works
Two distinct mechanics, both required for org-owned repos:
Indexing — paste a GitHub URL into the "Add Server" button on glama.ai (modal in the SPA, no public API). Glama then crawls the repo and indexes tools/schemas/annotations. This is how our 5 currently-indexed servers (autotask-mcp, hudu-mcp, qbo-mcp, rocketcyber-mcp, xero-mcp) got listed.
Claiming — add a glama.json to repo root, then re-trigger the claim flow on Glama. From the Glama blog:
For servers hosted under an organization, you must use the glama.json file. ... If after adding a glama.json you are still not able to claim the ownership of the MCP Server page, you should ensure that you've added your GitHub username.
Personal GitHub auth alone does not work for org repos — the file is mandatory. The exact verification Glama performs on the claim flow isn't documented; inferred to be: claimer is authenticated GitHub user + listed in maintainers + has repo write access (implicit via org membership). None of our 5 indexed servers are currently claimed — their attribute lists show only hosting:*, no author:claimed or author:official.
Canonical glama.json (matches autotask-mcp today)
{
"$schema": "https://glama.ai/mcp/schemas/server.json",
"maintainers": ["asachs01"]
}
The Glama schema currently only defines one property: maintainers (array of GitHub usernames). Nothing else to template.
Current state (2026-05-20)
Scope: 44 in-scope *-mcp repos (excluding afkbot-mcp and brain-mcp per maintainer decision — internal/experimental, not for public directory).
Already have glama.json (20)
action1-mcp, atera-mcp, autotask-mcp, blumira-mcp, connectwise-automate-mcp, connectwise-manage-mcp, datto-rmm-mcp, domotz-mcp, halopsa-mcp, hudu-mcp, huntress-mcp, itglue-mcp, liongard-mcp, ninjaone-mcp, qbo-mcp, rocketcyber-mcp, salesbuildr-mcp, superops-mcp, syncro-mcp, xero-mcp
Missing glama.json — need fleet PR (24)
abnormal-mcp, avanan-mcp, blackpoint-mcp, cipp-mcp, crewhu-mcp, datto-bcdr-mcp, datto-saas-protection-mcp, immybot-mcp, iqms-mcp, ironscales-mcp, kaseya-bms-mcp, kaseya-vsa-mcp, knowbe4-mcp, mimecast-mcp, pax8-mcp, proofpoint-mcp, rootly-mcp, sentinelone-mcp, sherweb-mcp, spamtitan-mcp, spanning-mcp, threatlocker-mcp, timezest-mcp, unitrends-mcp
Currently indexed (5)
autotask-mcp, hudu-mcp, qbo-mcp, rocketcyber-mcp, xero-mcp — all already have glama.json. Unclaimed.
Currently unindexed (39)
Everything in scope except the 5 above. Needs manual "Add Server" submission per repo.
Proposed sequence — pilot first, then fleet
- Pilot the claim flow on one already-indexed repo (e.g.
autotask-mcp). Sign into Glama as asachs01, trigger the claim flow against the existing glama.json. Confirm the listing flips to author:claimed and the edit affordances appear.
- If step 1 succeeds: fleet PR adding
glama.json to the 24 missing repos. One commit per repo (or one batched commit per repo via a small script). Identical 4-line file in each.
- Submit the 39 unindexed repos via Glama's "Add Server" UI. Each ~30 seconds. Could be browser-automated since we control the logged-in session, but 39 manual submissions is also fine.
- Trigger claim flow on each repo post-indexing — should be near-automatic for the 39 new submissions (they'll have
glama.json in place from day one), and a one-time sweep for the 4 already-indexed unclaimed ones (hudu, qbo, rocketcyber, xero).
- Update the daily digest's success criteria to also report claimed-vs-unclaimed counts once that distinction is meaningful.
Open questions
- Step 1 might fail if Glama's verification is stricter than the docs suggest (e.g. requires the GitHub identity be a public member of the org, or requires repo-level admin not just push). If so, we need to either go public on org membership, or escalate to Glama support@glama.ai for the org-claim path.
pax8-mcp in the missing list — Pax8 is a distributor we integrate with, not a vendor we own end-to-end. Worth a sanity check that listing it under wyre-technology org auspices is the right call vs. asking Pax8 to host their own MCP server.
Out of scope
Acceptance
- Daily stars-watcher digest reports
44 of 44 servers indexed by Glama with most/all claimed.
Companion to #16. Tracks Glama.ai coverage for the
*-mcpfleet. Surfaced by stars-watcher's daily digest — currently reporting 5 of 45*-mcprepos indexed by Glama.How Glama actually works
Two distinct mechanics, both required for org-owned repos:
Indexing — paste a GitHub URL into the "Add Server" button on glama.ai (modal in the SPA, no public API). Glama then crawls the repo and indexes tools/schemas/annotations. This is how our 5 currently-indexed servers (
autotask-mcp,hudu-mcp,qbo-mcp,rocketcyber-mcp,xero-mcp) got listed.Claiming — add a
glama.jsonto repo root, then re-trigger the claim flow on Glama. From the Glama blog:Personal GitHub auth alone does not work for org repos — the file is mandatory. The exact verification Glama performs on the claim flow isn't documented; inferred to be: claimer is authenticated GitHub user + listed in
maintainers+ has repo write access (implicit via org membership). None of our 5 indexed servers are currently claimed — their attribute lists show onlyhosting:*, noauthor:claimedorauthor:official.Canonical glama.json (matches
autotask-mcptoday){ "$schema": "https://glama.ai/mcp/schemas/server.json", "maintainers": ["asachs01"] }The Glama schema currently only defines one property:
maintainers(array of GitHub usernames). Nothing else to template.Current state (2026-05-20)
Scope: 44 in-scope
*-mcprepos (excludingafkbot-mcpandbrain-mcpper maintainer decision — internal/experimental, not for public directory).Already have
glama.json(20)action1-mcp,atera-mcp,autotask-mcp,blumira-mcp,connectwise-automate-mcp,connectwise-manage-mcp,datto-rmm-mcp,domotz-mcp,halopsa-mcp,hudu-mcp,huntress-mcp,itglue-mcp,liongard-mcp,ninjaone-mcp,qbo-mcp,rocketcyber-mcp,salesbuildr-mcp,superops-mcp,syncro-mcp,xero-mcpMissing
glama.json— need fleet PR (24)abnormal-mcp,avanan-mcp,blackpoint-mcp,cipp-mcp,crewhu-mcp,datto-bcdr-mcp,datto-saas-protection-mcp,immybot-mcp,iqms-mcp,ironscales-mcp,kaseya-bms-mcp,kaseya-vsa-mcp,knowbe4-mcp,mimecast-mcp,pax8-mcp,proofpoint-mcp,rootly-mcp,sentinelone-mcp,sherweb-mcp,spamtitan-mcp,spanning-mcp,threatlocker-mcp,timezest-mcp,unitrends-mcpCurrently indexed (5)
autotask-mcp,hudu-mcp,qbo-mcp,rocketcyber-mcp,xero-mcp— all already haveglama.json. Unclaimed.Currently unindexed (39)
Everything in scope except the 5 above. Needs manual "Add Server" submission per repo.
Proposed sequence — pilot first, then fleet
autotask-mcp). Sign into Glama asasachs01, trigger the claim flow against the existingglama.json. Confirm the listing flips toauthor:claimedand the edit affordances appear.glama.jsonto the 24 missing repos. One commit per repo (or one batched commit per repo via a small script). Identical 4-line file in each.glama.jsonin place from day one), and a one-time sweep for the 4 already-indexed unclaimed ones (hudu, qbo, rocketcyber, xero).Open questions
pax8-mcpin the missing list — Pax8 is a distributor we integrate with, not a vendor we own end-to-end. Worth a sanity check that listing it underwyre-technologyorg auspices is the right call vs. asking Pax8 to host their own MCP server.Out of scope
Acceptance
44 of 44 servers indexed by Glamawith most/all claimed.