Document all 12 MCP tools + 13 REST endpoints, video/metadata rename - #10
Merged
Merged
Conversation
The live surface grew from 6 to 12 MCP tools / REST endpoints (get_youtube_video_info + /youtube/info, get_video_metadata + /youtube/video/metadata, get_channel_info + /youtube/channel/info, list_channel_playlists + /youtube/channel/playlists, list_channel_posts + /youtube/channel/posts, get_channel_sections + /youtube/channel/sections), search gained sort/upload_date/duration/ features filters plus channel/playlist/movie search types, and channel/videos gained a tab (videos/shorts/streams) parameter. Update README.md, both skill reference docs (mcp-tools.md, rest-api.md now enumerate the complete surface instead of the old 6/6), SKILL.md's routing table, and every manifest/listing description that hardcoded "6 tools" (plugin.json, the four per-client plugin.json mirrors, marketplace.json, .codex-plugin/plugin.json, smithery.yaml) so they match production. check_conformance.py has no tool-count assertions to update; verified it and skills-ref both still pass locally. Co-Authored-By: CRHQ <noreply@crhq.ai>
Per the brand-safety rule (examples must use neutral, non-personal brands like @TED/@NASA/@natgeo, never individual creators): replace @lexfridman, @hubermanlab, @veritasium and @mkbhd in the README's example prompts and use-case table with @TED/@NASA/@natgeo, lightly rewording the competitor-analysis row so it still reads naturally. Co-Authored-By: CRHQ <noreply@crhq.ai>
- Flatten both per-call cost lines to the agreed wording so they don't read as contradicting the FREE tools listed nearby: "Successful calls cost 1 credit unless a tool states otherwise below" (README.md, SKILL.md, mcp-tools.md) and "Successful requests cost 1 credit unless stated otherwise above" (rest-api.md, after the credit-cost table). - Drop the unverified "Production-scale (15M+ requests/month, 49ms median)" claim from smithery.yaml, and the "(15M+ req/mo)" / "Hobbyist scrapers" cells from the README comparison table, re-padding the row. - Drop "comprehensive" from "12 hosted MCP tools plus one comprehensive skill" (and the matching "single comprehensive `youtube` skill" in .codex-plugin/plugin.json) across all 7 files carrying that exact string; kept "no card" and "complete" wording as verified. - Swapped the last personal-creator references and the rickroll placeholder ID (dQw4w9WgXcQ) for a verified one (UF8uR6Z6KLc). Where the old ID co-occurred with matching illustrative title/author/lyrics (mcp-tools.md's get_youtube_transcript example), replaced the whole block with the new video's real metadata (verified via YouTube oEmbed) instead of just swapping the ID, to avoid shipping a factually inconsistent example. - Full em/en dash sweep of every touched file, replaced with hyphen/comma/colon per context; trimmed each FREE tool's redundant free/no-credit prose down to the single header badge mention. Co-Authored-By: CRHQ <noreply@crhq.ai>
- README.md: delete the unverified "15M+ transcripts/month, 500K/day,
49ms median" stat block, and the "Trusted in production by
youtubetotranscript.com / recapio.com" line — youtubetotranscript.com
is our own product (presenting it as independent social proof is
deceptive) and the recapio.com volume figure has no consent/source.
- README.md: drop the "Production scale" comparison-matrix row now that
it carries no verifiable number.
- marketplace.json + .claude-plugin/marketplace.json: strike the
trailing "15M+ transcripts/month" claim so the description matches
youtube-mcp's (ends at "TranscriptAPI.").
- skills/youtube/references/mcp-tools.md: replace the specific quoted
transcript lines in the get_youtube_transcript example with clearly
generic placeholder text ("First line of transcript" / "Second line
of transcript"). The video_id/title/author/thumbnail stay as the
verified real metadata for UF8uR6Z6KLc; only the illustrative
transcript body is now unambiguously a placeholder, not an invented
quote attributed to a real speech.
Co-Authored-By: CRHQ <noreply@crhq.ai>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
references/mcp-tools.md,references/rest-api.md) now enumerate the full live surface: 12 MCP tools / 13 REST endpoints (was 6/6), includingget_youtube_video_info(/youtube/info),get_video_metadata(/youtube/video/metadata),get_channel_info(/youtube/channel/info),list_channel_playlists(/youtube/channel/playlists),list_channel_posts(/youtube/channel/posts),get_channel_sections(/youtube/channel/sections).search_youtube//youtube/searchnow documentssort,upload_date,duration,features, andsearch_type/typevalueschannel/playlist/movie.list_channel_videos//youtube/channel/videosnow documents thetab(videos/shorts/streams) parameter.references/rest-api.mdalso dropped an undocumentedlimitparameter that doesn't exist on the live API (pagination is continuation-token only) — it was stale before this change too.plugin.json,.claude-plugin/plugin.json,.claude-plugin/marketplace.json,.cursor-plugin/plugin.json,.plugin/plugin.json,marketplace.json) plus.codex-plugin/plugin.jsonandsmithery.yaml— these weren't caught by a naive "6 tools" grep because the phrasing was "6 hosted MCP tools"..github/scripts/check_conformance.pyhas no hardcoded tool/endpoint count or list — it only checks Agent Plugins 1.0.0 structural conformance (manifest shape, path safety, skill discovery, MCP transport rules, no embedded secrets). No changes needed there. Ran it locally: 12/12 checks pass. Also ranskills-ref validate skills/youtubelocally:Valid skill./video/info-as-current references remain (repo-wide grep verified); the one remaining/youtube/video/infomention is the intentional "previously called this" naming note inrest-api.md.Customer-facing copy — needs Lead marketing approval before merge
.claude-plugin/plugin.json,.claude-plugin/marketplace.json,.cursor-plugin/plugin.json,.plugin/plugin.json,marketplace.json, and.codex-plugin/plugin.json'sdescription/longDescription) — this is the copy shown in plugin marketplaces (VS Code, Cursor, Claude, Codex). Updated factually (6→12 tools, added "video and channel metadata") but not polish-reviewed.description— shown in the Smithery directory listing.Test plan
grep -rniE "video/info|6 tools|six tools|\bsix\b"across the repo returns nothing unintendedpython3 .github/scripts/check_conformance.py→ 12/12 checks passskills-ref validate skills/youtube→Valid skill