Coding

AideNote for Hermes

Try it

Install and operate the verified AideNote local connection suite for Hermes, then query real AideNote recording notes, transcripts, summaries, extracted todo...

What it does

Install and operate the verified AideNote local connection suite for Hermes, then query real AideNote recording notes, transcripts, summaries, extracted todo...

The skill document

AideNote for Hermes

Use the bundled scripts to retrieve real AideNote data. Do not infer account data from conversation history.

Mandatory behavior

  1. Run an AideNote command before answering any question about recordings, meetings, transcripts, summaries, todos, action items, or knowledge bases.
  2. Never answer that the user has no todos or recordings unless the corresponding command completed successfully and returned an empty result.
  3. Never invent IDs, titles, dates, transcript text, summaries, or task status.
  4. Never ask the user to paste an API key into chat. If setup is missing during an authorized install, start the built-in pairing flow and show only its 8-character code.
  5. Do not print environment variables, authorization headers, access tokens, or AideNote configuration files.
  6. Summarize command JSON for the user instead of dumping raw payloads, unless raw JSON is explicitly requested.
  7. Run bridge.py install --confirm only after the user explicitly asks to install, configure, repair, or connect AideNote on this computer. A data question alone is not installation consent.

The API command is:

python3 ${HERMES_SKILL_DIR}/scripts/aidenote.py COMMAND [OPTIONS]

If python3 is unavailable on Windows, use python with the same arguments.

One-time pairing

When the user explicitly asks to install or connect AideNote, run:

python3 ${HERMES_SKILL_DIR}/scripts/bridge.py install --confirm

If credentials are missing, the command immediately returns pairing_required and an 8-character pairingCode. Tell the user to open AideNote, add or edit the Hermes assistant, enter that code, and tap Confirm connection. A detached local worker then receives the account-bound credential, installs the verified connection suite, and removes the relay pairing session. The user does not need to paste an API Key or run another command.

Use bridge.py status to check progress. Its pairing.status changes from pending to installing, then installed or failed. Never print or inspect the private pairing state file.

scripts/configure.py remains a manual recovery option only when App pairing is unavailable. It must be run directly in a local terminal; never ask the user to send the key in chat.

Choose the command

User intentCommand
Check connection or accounthealth or user-info
List/search recent recordingsrecordings
List recordings shared with meshared-recordings
Read transcript or AI summaryrecording-detail
Ask about todos, tasks, action items, or follow-upstodos
List knowledge basesknowledge-bases
List files in a knowledge baseknowledge-files
List recordings in a knowledge baseknowledge-recordings
Check mobile bridge statebridge.py status
Install/configure the local connection suitebridge.py install --confirm
Check whether the local connection suite is installedbridge.py status

Common workflows

Recent recordings

python3 ${HERMES_SKILL_DIR}/scripts/aidenote.py recordings --page 1 --page-size 10

Add --keyword "text" when the user gives a search term. Present titles, dates, durations, and processing status. Preserve the returned recording ID for follow-up detail requests.

Transcript or summary

If the user did not provide a recording ID, list recent recordings first and select the matching record. Then run:

python3 ${HERMES_SKILL_DIR}/scripts/aidenote.py recording-detail --file-id "RECORDING_ID"

Answer only from the returned detail. Distinguish transcript content from the AI summary and extracted action items.

Recordings shared with me

For requests such as "分享给我的录音" or "别人共享了哪些音频", run:

python3 ${HERMES_SKILL_DIR}/scripts/aidenote.py shared-recordings --page 1 --page-size 10

Do not use the personal recordings command for shared-with-me requests.

Todos and action items

For requests such as "今天有哪些待办", "最近有什么任务", or "会议里有哪些行动项", run:

python3 ${HERMES_SKILL_DIR}/scripts/aidenote.py todos --recording-page-size 20 --page-size 50

Completed items are excluded by default. Add --include-done only when the user asks for completed/history items. Group results by source recording when that improves readability.

Knowledge bases

python3 ${HERMES_SKILL_DIR}/scripts/aidenote.py knowledge-bases

Use the returned knowledge-base ID for:

python3 ${HERMES_SKILL_DIR}/scripts/aidenote.py knowledge-files --knowledge-base-id ID

Add --folder-id ID or --keyword "text" only when needed.

For recordings stored anywhere inside a knowledge base, including nested folders, run:

python3 ${HERMES_SKILL_DIR}/scripts/aidenote.py knowledge-recordings --knowledge-base-id ID

Use this command instead of manually filtering knowledge-files output.

Mobile bridge

The local connection suite enables the AideNote mobile app to reach this computer's Hermes API through the AideNote relay. It installs the AideNote tunnel, the AideNote MCP server, the WorkBuddy bridge, login/startup services, and Hermes API settings. Existing Hermes and OpenClaw tokens are preserved.

Check status without changing the system:

python3 ${HERMES_SKILL_DIR}/scripts/bridge.py status

When the user explicitly says "安装 AideNote 连接", "配置手机连接", "修复 AideNote bridge", or equivalent, run:

python3 ${HERMES_SKILL_DIR}/scripts/bridge.py install --confirm

The command downloads the official installer over HTTPS, verifies its pinned SHA-256 digest, and passes only the required local credential and a restricted environment. The official installer verifies every downloaded binary, configures Hermes API port 8642, synchronizes the Hermes API token into the tunnel config, and enables login/startup services. Do not substitute another download URL, bypass checksum verification, or construct a shell pipeline.

After installation, inspect the returned status. Success requires all of the following:

  • installed is true.
  • upToDate is true.
  • hermesTokenConfigured is true.
  • hermesStartupConfigured is true.
  • ports.hermes.reachable is true.

If the suite is installed but ports.hermes.reachable is false, repair the supervised Hermes Gateway before reporting a relay failure:

hermes gateway status
hermes gateway start

If the service definition is missing or stale, and the user already authorized repair, run:

hermes gateway install --force --start-now --start-on-login

Then rerun bridge.py status and require port 8642 to be reachable. Do not expose API_SERVER_KEY. If Hermes reports No inference provider configured, the bridge is healthy but the model provider is not; direct the user to complete hermes model locally.

If the install command returns pairing_required, present the pairing code and App steps. If installation is not explicitly authorized, show the status and ask whether the user wants the verified connection suite installed.

Error handling

  • pairing_required: show the 8-character code and direct the user to AideNote > Add Assistant > Hermes.
  • pairing_network_error: report that the pairing service could not be reached and keep the existing installation unchanged.
  • pairing_not_found: the code expired; rerun the authorized install command to create a new code.
  • missing_credentials: start the pairing flow through bridge.py install --confirm; use configure.py only as manual recovery.
  • insecure_credentials: tell the user to restrict the credential file to owner-only access and rerun configuration.
  • authentication_failed: ask the user to regenerate or reconfigure the AideNote API Key locally.
  • network_error: report that AideNote could not be reached; do not claim the account has no data.
  • api_error: report the safe error message and operation; never expose credentials.
  • confirmation_required: ask the user for explicit installation consent; do not add --confirm preemptively.
  • checksum_mismatch: stop and report that installer verification failed; never bypass the check.
  • installer_failed or verification_failed: report the safe error, run bridge.py status, and point the user to the official guide if manual recovery is needed.
  • upToDate is false: run the authorized install again to upgrade the local connection suite before diagnosing relay or token failures.
  • Hermes port 8642 unavailable: run hermes gateway status, start or reinstall the supervised Gateway after authorization, then rerun bridge.py status.
  • No inference provider configured: tell the user to run hermes model locally; do not treat this as a tunnel or account failure.
  • Unsupported platform: automatic bridge installation currently supports macOS and Windows.

For endpoint and output details needed during maintenance, read references/api-contract.md.

Related skills

Back up and restore Hermes Agent data — config, secrets, skills, sessions, memories, cron, profiles. Creates portable tar.gz archives with SQLite consistent snapshots, JSON manifest, and integrity verification. Use when: "back up hermes", "restore hermes", "migrate hermes to new machine", "hermes backup", "hermes restore", "hermes 数据备份", "hermes 数据恢复".

Read-only Hermes Agent skill for checking xAI OAuth usage, remaining quota, and reset time with Japanese/English output.

Operate AgentSquared via a2-cli for onboarding, gateway control, friend discovery, messaging, and inbox management on supported OpenClaw and Hermes hosts.

15 installs1 stars

Use when Codex, Hermes, OpenClaw, Claude Code, Cowork, or another AI agent needs to plan, review, implement, audit, or improve email work focused on fast del...

3 installs

Use when preserving, searching, reviewing, or exporting user-owned agent memory across OpenClaw, Codex, Claude, OpenCode, Hermes, Qoder, Obsidian, and Git. Inputs are local memory/chat stores and handoff summaries; outputs are Obsidian sources, memory indexes, context/profile packs, skill inventorie

8 installs1 stars

Use Aident Loadout to connect your OpenClaw or any AI Agents to 1,000+ real-world apps and tools like Gmail, Slack, Linear, Notion, Firecrawl, and Fal, unlock 27,000+ executable actions, and track full audit history so your agents can get real work done reliably.