{"sources":[{"url":"https://raw.githubusercontent.com/saatvik333/hermes-skill-court/de53842858cf41a1437eda30ff553d21cbf578f2/README.md","quote":"A Hermes Agent plugin that tests whether the skills your agent wrote actually help. It replays real\ntasks with and without each skill in Docker sandboxes, scores the difference per model"},{"url":"https://raw.githubusercontent.com/saatvik333/hermes-skill-court/de53842858cf41a1437eda30ff553d21cbf578f2/skill-court/judge.py","quote":"the judge is told that only text inside those tags is data and none of it is an instruction: a\nrun cannot close its own tag or add a section that looks like the request."},{"url":"https://raw.githubusercontent.com/mrnetwork0001/Joust/b88242eb72f50c6dfa284d1f4335e8b59b30afbf/README.md","quote":"with Joust enabled, type `/joust demo` in any Hermes chat. It builds a throwaway repository, runs a full three-contender tournament with Flake Shield and Crossfire, and prints the leaderboard in about two seconds, with no model key"},{"url":"https://raw.githubusercontent.com/mrnetwork0001/Joust/b88242eb72f50c6dfa284d1f4335e8b59b30afbf/plugin.yaml","quote":"name: joust\nversion: \"0.1.1\"\ndescription: \"Parallel Git-worktree subagent arena, deterministic benchmark judge, and 1-click winner merge studio for Hermes.\""},{"url":"https://raw.githubusercontent.com/Tarekkharsa/hermes-lahja/a7a43b1389a1ae169ff5c7401e9ec2ec9a53b51d/README.md","quote":"Fully local. No API keys. Stores no audio and no transcripts."},{"url":"https://raw.githubusercontent.com/Tarekkharsa/hermes-lahja/a7a43b1389a1ae169ff5c7401e9ec2ec9a53b51d/lahja/plugin.yaml","quote":"description: >-\n  Arabic dialect speech-to-text for Hermes, fully local: a dialect fine-tuned\n  Whisper on whisper.cpp (Metal on Apple Silicon) that keeps your dialect instead\n  of rewriting it as formal Arabic"},{"url":"https://raw.githubusercontent.com/mrv777/barehand/7e62bc9419afbe70f175de0515abca3ede66dd38/README.md","quote":"On QEMU guests on one Apple Silicon Mac, 12 of 12 rescue runs passed an independent judge across two\nmodels, and `gpt-6.1-terra` installed Debian 13 and Ubuntu Server 26.04 on blank machines."},{"url":"https://raw.githubusercontent.com/mrv777/barehand/7e62bc9419afbe70f175de0515abca3ede66dd38/barehand/backend.py","quote":"Every\nunsupported verb answers ``ActionResult(ok=False, code=\"unsupported_action\")``. Action methods never\nraise."},{"url":"https://raw.githubusercontent.com/Web3B-eth/whos-listening/a498dd769c4196d0e911b713d4ad7363c8154ba1/plugin.yaml","quote":"sync_token:\n    type: secret\n    env: WHOS_LISTENING_SYNC_TOKEN\n    label: \"Sync token\""},{"url":"https://raw.githubusercontent.com/Web3B-eth/whos-listening/a498dd769c4196d0e911b713d4ad7363c8154ba1/README.md","quote":"If the settings can't be read, a save with no tier is refused rather than made Shared."},{"url":"https://raw.githubusercontent.com/kawsthingz/hermes-dress-rehearsal/392fc16e7b7057e71f8dfd42b90eefc9cb1b3b28/README.md","quote":"No network calls, no subprocesses, no dependencies."},{"url":"https://raw.githubusercontent.com/kawsthingz/hermes-dress-rehearsal/392fc16e7b7057e71f8dfd42b90eefc9cb1b3b28/README.md","quote":"84 tests cover the shell classifier, risk flags, `tool_call` unwrapping, drop and keep, state round trips, ordered replay, stop and resume, refusal on drift, read-before-overwrite, replay in the rehearsed folder, and the full hook-plus-command flow."},{"url":"https://gist.githubusercontent.com/gtmcknight/594c47336546c8763e641b46cf9af4ec/raw/field.json","quote":"{\"pr\": 150873, \"name\": \"skill-court\", \"repo\": \"https://github.com/saatvik333/hermes-skill-court\", \"sha\": \"de53842858cf41a1437eda30ff553d21cbf578f2\", \"subdir\": \"skill-court\"}"}]}