GitHub profilePlugins by Sahil-SS9
10 plugins by Sahil-SS9 in the catalog, most-starred first.
Sahil-SS9hermes-project-stewardshipDurable project ownership and bounded initiative management for Hermes fleets. Disclosure — a `pre_llm_call` hook adds a project's saved rules, decisions and facts to turns of bots assigned to that project; objective checks run allowlisted local programs only at autonomy level 3+ and GitHub objectives shell out to your own `gh api`; it creates and edits Hermes Kanban tasks through Hermes' internal kanban modules (title/body/tenant written directly to kanban.db), reads other Hermes profiles' project registries (projects.db, read-only, profile names and project counts only) to hint where existing projects live during onboarding, keeps its own SQLite DB under `<Hermes home>/plugin-data/hermes-dockyard/`, and its optional HTTP API (`/stewardship/v1`, 127.0.0.1) stays closed until you set a bearer token.Communityautomation82Sahil-SS9prompt-optimizerModel-aware prompt rewriting with quality scoring, analytics, and coaching. Disclosure — the default auto mode rewrites every user message of 5+ words through your configured model (one extra LLM call per turn) before the agent sees it.Communitytools34Sahil-SS9toolariaRescues oversized MCP and web tool results before they flood context.Communitytools22Sahil-SS9memlockReasserts pinned standing instructions after context compaction. Disclosure — scope=global pins apply to every session sharing the same HERMES_HOME; memlock_setup.py edits config.yaml and .claude/settings.json (with backups) only when you run it.Communitytools18Sahil-SS9severianLocal-first long-term memory for Hermes agents: a canonical append-only store (SQLite or PostgreSQL+pgvector) with five-dimensional tenant/profile/collection/agent/session scope, hybrid BM25+vector retrieval with a re-ranker, memories that cite the observations they came from, per-turn user facts, real dates next to relative times, truth resolution when a newer fact replaces an older one, and a conflict note when two returned memories disagree. Fully offline by default; Qwen3-Embedding-0.6B semantic embeddings opt-in, locally or in a user-owned sidecar. On OmniMemEval's public long-term-memory leaderboard (its own models, prompts and full public sets) it scores 90.6 LongMemEval (1st), 87.3 LoCoMo (2nd) and 71.6 BEAM 100K (2nd, within noise of the lead); it re-ran LCM-X to check the board's records, and per-question results, a strict second marker and scripts are published in the repo. Disclosure — stores every conversation turn in one SQLite database under the Hermes root (`<hermes root>/severian`) shared by all profiles with rows scoped per profile, or in the PostgreSQL database named by `SEVERIAN_DSN`; the opt-in Qwen/Granite embeddings download a model from Hugging Face on first use, and the opt-in embedding sidecar sends memory text to the URL you configure.Communitymemory5Sahil-SS9spectatorHarness-neutral live session sharing, summaries, and branch queries. Disclosure — reads state.db read-only and exposes masked transcripts to invite-token guests on the dashboard; host-approved guest suggestions are injected via ctx.inject_message.Communitytools5Sahil-SS9agent-reelTurns your own agent history into a comic (Hermes, Claude Code, Codex, OpenCode or any JSONL log). It reads your sessions, tool calls and kanban cards read-only, finds a story with a problem, a struggle and a result, and writes a chapter script where every line and caption carries the ids of the messages that state it — a claim the cited messages do not support is refused, not printed. A series ledger remembers what each chapter told, carries one open thread into the next, and refuses repeats; an idea deck of 16.9 billion frames keeps each chapter's world, villain and genre different. It then draws the episode with your own image route (your harness's image tool, a local ComfyUI, or your own API key) in one of 55 art styles, letters it with SIL OFL fonts, and exports pages, panels, X/LinkedIn/Discord sizes and a PDF. Disclosure — reads your Hermes history read-only (no file created, renamed, truncated or deleted, proven with strace in scripts/readonly_proof.sh), reads only the profile it runs as unless --all-profiles is given, redacts deterministically by default (the local model redaction pass is opt-in via --llm-url), and sends redacted text to Hermes's own model for writing and truth checks, and panel prompts to the image route you choose; writes plans, renders, a series ledger and character kits under the plugin data directory; no telemetry, no auto-posting, and no key at all is needed for the host or ComfyUI routes.CommunitygeneralSahil-SS9agent-run-debuggerLocal-first evidence debugger for failed, stalled or expensive agent runs: import one explicitly selected trace and get a deterministic timeline, observed failure points, ranked hypotheses that cite their evidence, missing-evidence gaps and an inert replay of the recorded run. Disclosure — imports a trace file explicitly named by you or by the model through the rundebug_import tool and writes only under the active profile's plugin-data directory, with redaction before the store, the journal and exports; replay executes nothing (no subprocess, no network connection, no live database); no telemetry, no background capture, no stored credentials, no third-party network calls and no self-updater; the export command is operator-CLI-only and is not exposed through tool, slash or gateway routes.CommunitytoolsSahil-SS9hermes-agent-mixerA mixing desk for how much a coding agent may change: four enforced dials (edits, write-scope globs, per-turn line budget, new-dependency policy) with /mixer, live presets, a Desktop dial panel in the composer, a web-dashboard saved-defaults bar and a steering-console page. Also drives Claude Code, Codex and Gemini CLI through the agent-mixer hook command. Disclosure — enforces edit/scope/line-budget/new-dependency dials in pre_tool_call (blocks, or escalates "ask" to Hermes approval, which /yolo skips) and adds a one-line rules context via pre_llm_call when a dial is off its default; per-session dial state is stored under HERMES_HOME/plugin-data/hermes-agent-mixer/sessions/, and the Desktop panel and dashboard bar write the plugin's own saved defaults into config.yaml (plugins.entries.hermes-agent-mixer.settings); the separately installed agent-mixer CLI keeps its own state in ~/.config/agent-mixer and ~/.local/state/agent-mixer, writes hook entries into ~/.claude/settings.json, ~/.claude/commands/mixer.md, ~/.codex/hooks.json or ~/.gemini/settings.json only when run as `agent-mixer install <harness> --yes`, and `agent-mixer ui` serves a token-guarded console on 127.0.0.1:8377.CommunitygeneralSahil-SS9modernisation-engineControlled migration engine: maps an existing codebase, turns the behaviour that must survive into reviewable contracts, prepares a separate candidate and records old/new evidence before any production change. Disclosure — 27 tools plus a `hermes modernise` CLI and `/modernise` slash command; `assess`, `plan`, `record` and `recipe` read source trees and files at absolute paths the agent supplies (refusing symlinks, `.env`, `auth.json`, `credentials.json` and SSH keys) and never import or run repository code; `recipe` creates a new directory at an agent-chosen absolute path holding a copy of the source with fixed metadata edits; job records live under the plugin's data directory; only after you set `operator_config` and sign a one-use Ed25519 owner grant do `run`, `cancel` and `recover` invoke the local `docker` CLI, outside Hermes's command-approval prompt, to run, inspect and remove rootless, network-disabled, read-only, resource-limited containers from digest-pinned images you allow-list (never pulled); no network calls, no telemetry, no self-updater.Communitytools
