agentleFS
NVIDIAKnown publisherNVIDIA/nemoclaw-community/examples/recipes/nvidia/agentic-ai-learning-path/agents/hermes/skills/slack-channel-finder/SKILL.md

slack-channel-finder

Skill142 starsChanged 3 months agoApache-2.0View on GitHub

Discover Slack channels by topic, team, or domain and determine the channel ID.

The file

We split this skill into the rules it gives an agent

10 rules. About 0 are common, 7 are distinctive, and 3 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈7

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 3

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Use this skill to discover Slack channels ID by name or topic

    line 8
  2. Distinctive:

    Do NOT use this skill when the user has already provided a slack channel ID

    line 12
  3. Distinctive:

    Avoid writing custom curl or python commands

    line 17
  4. Distinctive:

    Only customize if the existing scrips fail, and in that case follow their authorization patterns closely

    line 18
  5. Distinctive:

    Check for scripts at:

    line 26
  6. Distinctive:

    If the user is asking to summarize or inspect a specific Slack channel, and they likely know the channel already but did not provide the …

    line 49
  7. Distinctive:

    If the user is asking you to help find which Slack channel is relevant to a topic, team, or project, use the discovery flow below instead …

    line 52
  8. Very specific:

    When you need to understand a specific channel — what it's for, who's active, what they're discussing — use describe_slack_channel.py:

    line 142
  9. Very specific:

    When a message has a high reply_count and you need the actual discussion content, use --replies on describe_slack_channel.py or directly …

    line 226
  10. Very specific:

    If the user's goal goes beyond discovery ("tell me what the X team is working on"), once channels are identified, hand off to …

    line 238

≈ “Common” is an estimate: at least 1 in 5 of the 364,980 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

slack-channel-finder

Use this skill to discover Slack channels ID by name or topic

When to use

Do NOT use this skill when the user has already provided a slack channel ID

Instructions

  • Because of the sandbox, the best path is to use the provided helper scripts
  • Avoid writing custom curl or python commands
  • Only customize if the existing scrips fail, and in that case follow their authorization patterns closely
  • Be aware that slack access is through a Slack bot, whose token might have certain scopes or might be missing scopes

Scripts

Check for scripts at:

/sandbox/.hermes-data/skills/slack-channel-finder/scripts/

or at

/sandbox/.hermes/skills/slack-channel-finder/scripts
ScriptPurpose
find_channel.pySearch and rank channels by query across the workspace
list_accessible_channels.pyList channels (bot-member or workspace-wide)
describe_slack_channel.pyDeep-describe a single channel with layered signals

Procedure

1. Distinguish summary requests from discovery requests

  • If the user is asking to summarize or inspect a specific Slack channel, and they likely know the channel already but did not provide the exact name or ID, ask them for the channel name.
  • If the user is asking you to help find which Slack channel is relevant to a topic, team, or project, use the discovery flow below instead of asking for a channel name first.

2. If you are given a channel name, start with list_accessible_channels.py and try to match channel name to ID

/usr/bin/python3 .../list_accessible_channels.py

3. If you are unclear on what channel the user is talking about, ask them for the channel name

4. If the user wants help finding a channel, attempt to find one based on the user's topic or query

For topic or team queries, use find_channel.py — it searches all discoverable public channels (not just bot-member channels) and returns scored matches:

/usr/bin/python3 /sandbox/.hermes-data/skills/slack-channel-finder/scripts/find_channel.py \
  --query "nemoclaw inference" --top 5

Output:

{
  "ok": true,
  "query": "nemoclaw inference",
  "query_tokens": ["nemoclaw", "inference"],
  "total_searched": 78,
  "count": 2,
  "discovery_mode": "workspace",
  "results": [
    {
      "channel_id": "C0ASZUN3L5D",
      "name": "lopp-nemoclaw-staging",
      "is_member": true,
      "num_members": 8,
      "topic": "",
      "purpose": "This is a channel to discuss NemoClaw technical updates...",
      "score": 9,
      "match_reasons": ["name:nemoclaw", "purpose:nemoclaw"]
    }
  ]
}

Scoring weights: name token match = 3 pts, purpose match = 2 pts, topic match = 1 pt.

The is_member flag tells you whether the bot is in the channel — full history and thread signals are available only for member channels.

Options:

FlagDescription
--query TEXTRequired. Matched against name, topic, purpose.
--top NMax results (default 5)
--member-onlyRestrict to bot-member channels only
--min-score NMinimum score to include (default 1)

Information on other scripts

List all channels (when you need the full inventory)

For cases where you need the complete channel list rather than a scored search:

# Bot-member channels only (fast)
/usr/bin/python3 .../list_accessible_channels.py

# All public channels in the workspace
/usr/bin/python3 .../list_accessible_channels.py --all-public

Output: { "ok": true, "count": N, "channels": [...], "discovery_mode": "workspace" }

Each channel: {id, name, is_archived, is_private, is_member, num_members, topic, purpose, created}

For --all-public, is_member=false channels exist in the workspace but the bot hasn't been added — you can see name/topic/purpose but NOT read their history.

Options:

FlagDescription
--all-publicUse conversations.list for workspace-wide discovery
--include-archivedInclude archived channels
--types TYPESComma-separated types (default public_channel)

Describe a channel in depth

When you need to understand a specific channel — what it's for, who's active, what they're discussing — use describe_slack_channel.py:

# Full mode (name + topic + purpose + pins + bookmarks + recent history)
/usr/bin/python3 .../describe_slack_channel.py --channel-id C0ASZUN3L5D

# Fast mode (skips conversations.history — useful for breadth scans)
/usr/bin/python3 .../describe_slack_channel.py --channel-id C0ASZUN3L5D --no-history

# With thread content (expands reply threads for high-activity messages)
/usr/bin/python3 .../describe_slack_channel.py --channel-id C0ASZUN3L5D --replies

# With resolved user display names on top contributors
/usr/bin/python3 .../describe_slack_channel.py --channel-id C0ASZUN3L5D --resolve-users

Options:

FlagDescription
--channel-id IDRequired. Slack channel ID (e.g. C0ASZUN3L5D)
--history-limit NMax messages to fetch (default 50)
--no-historySkip conversations.history (faster, cheaper)
--no-pinsSkip pins.list
--no-bookmarksSkip bookmarks.list
--repliesFetch first few replies for threaded messages (reply_count > 0)
--replies-limit NMax replies per thread when --replies is set (default 5)
--resolve-usersResolve contributor user IDs to display names via users.info

Output structure:

{
  "ok": true,
  "channel_id": "C0ASZUN3L5D",
  "name": "lopp-nemoclaw-staging",
  "is_archived": false,
  "is_private": false,
  "num_members": 8,
  "signals": {
    "name_tokens": ["lopp", "nemoclaw", "staging"],
    "topic": "",
    "topic_stale": true,
    "purpose": "This is a channel to discuss NemoClaw...",
    "pinned_messages": [],
    "bookmarks": [],
    "recent_human_messages": [
      {
        "user": "UR0A4QL5N",
        "text": "<@U0AUN68FSNT> tell me a joke",
        "ts": "1777589708.784719",
        "thread_ts": "1777589708.784719",
        "reply_count": 2,
        "thread_messages": [...]
      }
    ],
    "top_contributors": [
      {"user_id": "U0887Q5UVV4", "message_count": 13, "display_name": "Scott Lopp"}
    ],
    "human_message_count": 22
  },
  "confidence": "medium"
}

The script does NOT produce a natural-language description. Synthesize one from the signals dict, weighting in this order:

  1. Pinned messages (often a charter or intro)
  2. Channel name tokens
  3. Topic and purpose (if not stale)
  4. Bookmarks
  5. Recent human message themes
  6. Top contributors

The confidence field (high, medium, low) reflects how many independent signals were available. For low-confidence channels, hedge ("appears to be about ...") or ask the user to confirm.

Other

Read thread content

When a message has a high reply_count and you need the actual discussion content, use --replies on describe_slack_channel.py or directly call conversations.replies:

# Via describe (expands all threads in the sampled history)
/usr/bin/python3 .../describe_slack_channel.py --channel-id [channel-id] --replies --replies-limit [n]

Chain into summarization if requested

If the user's goal goes beyond discovery ("tell me what the X team is working on"), once channels are identified, hand off to slack-channel-summarizer for each top channel. Cap at 5 channels per query; surface the ranking so the user can ask for more.

Common patterns

Find channels matching a topic:

/usr/bin/python3 .../find_channel.py --query "nemoclaw deployments"

See all public channels in the workspace:

/usr/bin/python3 .../list_accessible_channels.py --all-public

Understand a specific channel (fast, no history):

/usr/bin/python3 .../describe_slack_channel.py --channel-id C0ASZUN3L5D --no-history

Understand a channel with thread context and user names:

/usr/bin/python3 .../describe_slack_channel.py --channel-id C0ASZUN3L5D --replies --resolve-users

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.