v1.2.0: skills, slash commands, CI - #1
Merged
Merged
Conversation
Skills (thin by design; methodology stays server-side behind load_skill): - hypd-tools: stable tool-call invariants; init's output stays authoritative - hypd-getting-started: pre-auth onboarding, works before any tool can - hypd-overview: platforms + command table (the pre-install sales surface) Commands: audit, wasted-spend, negatives, keywords, feed-audit, meta-audit, creative, attribution, landing-page, report, prompts. Each is a thin entry point calling prompt_templates_run by key; keywords and meta-audit pick between two templates by context; every command degrades to connection guidance when unauthenticated. CI: JSON checks + claude plugin validate --strict + version-bump gate. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
…ble step 4 - Auth-fallback paragraph now lives ONLY in hypd-getting-started; the 11 commands and hypd-tools point to it in one line (was 12 verbatim copies) - Dashless-ID fact now lives only in hypd-tools - hypd-tools description trimmed to triggers (body carries the invariants) - 'Never imply otherwise' negation paired with positive instruction - Commands' step 4 completion criterion made checkable Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Member
Author
|
Pushed forge-review fixes: the auth-fallback block now has one home (hypd-getting-started) with the 11 commands + hypd-tools pointing to it, hypd-tools' description is trigger-only, and step-4 completion criteria are checkable. Validation still passes; always-on cost dropped ~702 → ~677 tok. Full review notes in HYP-905's folder (forge-improvements.md). |
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
1 participant
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.
Implements the spec settled in the HYP-905 grilling session.
3 skills — deliberately thin; all 12 methodology skills stay server-side behind
load_skill, routed byinit's resolver.hypd-toolscarries only stable invariants and defers toinitas authoritative (drift control). Fixes the visible "Skills: 0" in the plugin manager.11 commands — thin entry points wrapping
prompt_templates_runby key. Curated for platform coverage over raw Popular tags: Google Ads (audit, wasted-spend, negatives, keywords, report), Merchant Center (feed-audit), Meta (meta-audit, creative), GA4 (attribution), landing pages (landing-page), plus a/promptsbrowser over the full 87-template library.keywordsandmeta-auditeach select between two templates by context. Every command ends with the shared guide-don't-fail line for unauthenticated sessions.CI — JSON syntax,
claude plugin validate --strict, and a version-bump gate on PRs.Verified locally: validation passes; projected cost ~702 tok always-on across all components (shown to users pre-install).
Test before merge:
claude --plugin-dir .then try/hypd-ai-ads:auditand/hypd-ai-ads:prompts.Related: HYP-905
🤖 Generated with Claude Code