hermes-skills

Guides Cloudflare One Zero Trust and SASE work across Access, Gateway, WARP, Tunnel, Cloudflare WAN, DLP, CASB, device posture, and identity. Use when designing, configuring, troubleshooting, or reviewing Cloudflare One deployments. Retrieval-first: use current Cloudflare docs/API schemas instead of embedded product docs.

Use when scheduling recurring polls, read-only monitors, or background jobs that must survive shell session boundaries.

Discover, register, and maintain Hermes skills from non-default paths (plugin directories, external repos, shared vaults). Use when skills are missing from `skills_list()`, when the user mentions plugin skills, when symlinking skills, or when reconciling duplicate/overlapping skills between the default tree and external sources.

Operate the Antigravity CLI (agy): plugins, auth, sandbox.

Tune HERMES_STREAM_RETRIES for mid-stream reconnect resilience.

Manage parallel Hermes sessions in Telegram topics.

Run a startup checklist via a gateway:startup hook on every boot.

Discover, install, and manage agent skills from skills.sh.

Solar-aware 6-mode time-of-day theming with NOAA sunrise/sunset calculation

Solar-aware 6-mode time-of-day theming with NOAA sunrise/sunset calculation

Use sudo in terminal without interactive prompts via .env.

Control iTerm2 windows, tabs, panes, and sessions via Python API.

Tune agent.api_max_retries for fast fallback or resilience.

Break a concept into chunks and generate questions to explore. Used by Atlas+Scout primitives.

Configure MCP servers with env-var placeholders from .env.

Decide whether a concept is worth working on. Used by Atlas+Scout primitives.

Inject non-persisting system prompts via environment variables.

Integrate messaging platforms (Mattermost, Discord, Telegram, Slack, Matrix, SMS, email) with Hermes for bidirectional message delivery, media attachments, thread support, and fleet coordination.

Start Hindsight, Honcho (Cortex), and Holographic (OB1 + OBn) simultaneously. Includes verification and troubleshooting.

Capture user requests ("remember this") into all four canonical memory systems (Holographic, Honcho, Hindsight, OB1/OBn).

Recurring workflows to deepen and enrich Hermes' memory systems (L1, L2, L3, OBn).

Multi-source memory recall that synthesizes a cited answer with explicit abstention when evidence is insufficient. Use when the user needs a direct synthesized answer with citations rather than raw chunk dumps.

Re-runnable audit of the triple-layer memory system (L1 MEMORY.md/USER.md, L2 Hindsight :9177, L3 fact_store SQLite). Use when the user says "audit memory", "config memory", "memory-prune", or when MEMORY.md is near/exceeding budget. Bundles a one-shot script plus the procedure for what to do with the findings.

Session-end ritual for the L2 Hindsight layer. Drains durable session facts into the Hermes bank so they survive the L1 budget cull and become semantically searchable. Pairs with memory-sop-audit and the standard session-end ritual in AGENTS.md.

How and when to call fact_feedback on L3 facts. Without feedback the trust_score stays at the 0.5 default and `reason` / `contradict` queries have no signal. Skill exists to enforce the loop.

Fleet-wide Standard Operating Procedure for the triple memory system. Covers what goes where, when to write, when to read, and cross-layer hygiene. All fleet agents MUST follow this SOP.

Run at session end to reconcile tasks, update skills, and write to brain. Enforces closure discipline across the fleet.

Per-turn memory sync protocol using Open Brain (OB1) as the central knowledge graph. Every active session writes new facts/decisions to OB1 after each turn and pulls recent changes from other sessions before responding. This creates a live, shared knowledge graph that all fleet agents can read at any point during their work.

Coordinate approval-gated email campaigns.

Tune Hermes output density via display.compact and related settings.

Operate Linear as a shared agent task queue with receipts.

Route prompts by task-demand surface before spending tokens.

Detect and resolve duplicate skill names across the Hermes skill tree. Scans for name collisions between default skills and plugin skills, compares content, recommends canonical versions, and provides safe cleanup commands. Use when skills_list() shows unexpected duplicates, plugin skills shadow default skills, or context bloat from redundant skill loading is suspected.

Parse mixed source+requirement input to author and save a skill.