← all publishers

royashbrook

@royashbrook source repo

Report this profile Claim this profile

• 4 skills
• 0 followers
• 3 days ago last updated

1. ▌
   Hush · royashbrook bundle

   Use whenever an agent needs to STORE, GENERATE, or USE a secret such as an API token, signing key, or password. Hush has one hard rule: the agent never sees plaintext, so it never enters chat, transcripts, stdout, logs, the clipboard, or a temp file. A user-provided value enters once through a hidden prompt; a strong random value can be generated directly into the OS-backed store. Hush later injects it straight into a consumer through an environment variable or stdin, never prints it, and deliberately has no plaintext get command. Trigger on "store this token", "save this key", "add it to the keychain", "generate an operator/signing key", "use the X secret to call Y", or whenever an agent needs a credential to reach a service. macOS, Linux, and Windows backends are built in; the never-print contract is portable beyond them.
   0 installs
2. ▌
   SQL Spider · royashbrook bundle

   Build a deterministic dependency graph of SQL tables, views, procedures, functions, and triggers, then spider outward from a seed until the requested dependency frontier is closed. Use for "map this database", "what references table X", "what does proc Y touch", "find the closure around this table", "spider the schema", or "I cannot connect directly, give me queries I can run". The engine never opens a database connection: it emits READ-ONLY queries and ingests CSV results, so you, the agent, are the adapter. Run each query through whatever database access exists, feed the result back, and drive the loop to closure. SQL is parsed deterministically into an AST rather than guessed. T-SQL and SQLite ship; other dialects are adapters. Use `--graphify` to project a closed graph into graphify for community detection, query, path, and explain.
   0 installs
3. ▌
   Trustmebro · royashbrook bundle

   Insert verified links into a markdown or html post so a reader can READ and VERIFY it (hand them the source to check a claim, don't ask them to take your word; not a proof of truth). Three kinds: EXPLAIN a term a non-specialist would not know, SUBSTANTIATE a claim about an external thing, or (rarely) SHOW code via a sha-pinned permalink. Every link must resolve AND actually mean the thing in context (a live-but-wrong link is the worst output). NOT a fact-check of the document. Triggers: "cite this post", "add citations", "explain the jargon", "back up the claims", "make this post easier to follow". Sweeps jargon aggressively for a technical-generalist reader, auto-inserts verified links, then lists what it linked and invites you to add / remove / go more aggressive in one pass. Review the diff.
   0 installs
4. ▌
   Waterbear · royashbrook bundle

   Use when you want a Claude Code (or terminal AI) session to survive crashes, patches, quits, and reboots and come back on its own, when an agent is told to "waterbear yourself" / "make yourself always-on / unkillable / permanent", or when you want a respawn to RESUME the same conversation instead of starting blank. Most often an agent runs it on ITSELF to persist the current conversation. macOS/launchd reference implementation, adaptable to other OSes.
   0 installs