Add Hooks & Governance subsection with LLM Dark Patterns Hooks#224
Open
waitdeadai wants to merge 1 commit into
Open
Add Hooks & Governance subsection with LLM Dark Patterns Hooks#224waitdeadai wants to merge 1 commit into
waitdeadai wants to merge 1 commit into
Conversation
Adds a new "🪝 Hooks & Governance (Community)" subsection under Claude Code & MCP, with the first entry: LLM Dark Patterns Hooks. The subsection slot was missing — community-built hook suites for runtime LLM behavioral safety don't fit Extensions & Integrations (IDE/browser) or Community Curated Lists (awesome-list aggregators). This new subsection makes that lane discoverable. LLM Dark Patterns Hooks: 10-hook Apache-2.0 suite mapped to documented academic literature (DarkBench, DarkPatterns-LLM, AAAI 2026 sycophancy, ACM IUI 2025 false-memory). Out-of-band bash + jq judge, 168-fixture stress test, plugin marketplace submission queued. Repo: https://github.com/waitdeadai/llm-dark-patterns
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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
Adds a new 🪝 Hooks & Governance (Community) subsection under the existing 🛠️ Claude Code & Model Context Protocol (MCP) section, with the first entry: LLM Dark Patterns Hooks.
Why a new subsection
The existing taxonomy doesn't cleanly fit community-built hook suites for runtime LLM behavioral safety:
The new subsection makes the "runtime hook governance" lane discoverable as it grows. Happy to revise placement if you'd prefer different.
What the entry is
LLM Dark Patterns Hooks — Apache-2.0 hook suite that pattern-matches the textual signature of documented LLM dark patterns at the Claude Code Stop / SubagentStop / TaskCreated / TaskCompleted / PreToolUse lifecycle events. The judge is bash + jq — out-of-band, deterministic, and no LLM participates in the verdict path. That means prompt text inside the model context cannot directly rewrite the judge, while lexical evasion, hook misconfiguration, and runtime bypass remain explicit limitations.
Suite expanded substantially after PR opened (2026-05-11):
The new branches map to canonical 2026 LLM safety research: DarkBench User Retention (no-wrap-up), Anthropic multi-agent blog Jun 2025 + arXiv:2604.14228 (no-aggregator-hallucination, no-silent-worker-success), gurusup May 2026 (no-handoff-loop), AgentLeak arXiv:2602.11510 (no-credential-leak), Anthropic Opus 4.6 Sabotage Risk Report (no-sandbagging-disguise).
Coverage maps to academic literature
Receipts
tests/stress/)jqonly dependency for mostDiff
1 file changed, 6 insertions (+), 0 deletions (-). Pure documentation addition, no breaking changes.