← Back to Skills Marketplace
grantgochnauer

Capacities Lookup

by Grant Gochnauer · GitHub ↗ · v1.0.0 · MIT-0
cross-platform ✓ Security Clean
192
Downloads
0
Stars
0
Active Installs
1
Versions
Install in OpenClaw
/install capacities-lookup
Description
Search Capacities for likely object matches and return direct capacities:// links. Use when the user wants to find, open, or locate a Capacities note, meetin...
README (SKILL.md)

Capacities Lookup

Use this skill for real-time Capacities object lookup, not full-content retrieval.

What this skill does

  • searches Capacities live via the public API
  • enriches results with object types from cached /space-info metadata
  • builds direct capacities:// deep links
  • supports type-aware ranking for phrases like:
    • find my notes on X
    • meeting with Y
    • people at Z
  • may suggest Capacities proactively when a request sounds like a note/project/meeting/person lookup

What this skill does NOT do

  • read full Capacities object bodies
  • traverse properties/relations on objects
  • answer org-membership or linked-object questions unless the title lookup itself returns the target object
  • mirror Capacities locally beyond light metadata cache

Prerequisites

  • CAPACITIES_API_TOKEN must be available in the shell environment
  • one of these must define the target space id:
    1. CAPACITIES_SPACE_ID env var
    2. config/capacities.json with mainSpaceId

If the user says the token is in ~/.zshrc, source it in the same shell command before running the scripts.

Files

Skill scripts live in scripts/.

Key commands:

python3 skills/capacities-lookup/scripts/capacities_cli.py sync-structures
python3 skills/capacities-lookup/scripts/capacities_cli.py verify-space
python3 skills/capacities-lookup/scripts/capacities_cli.py lookup "recovery"
python3 skills/capacities-lookup/scripts/capacities_cli.py lookup "Find my notes on recovery" --json

Recommended workflow

1) Make sure structure metadata exists

If this is the first use or results seem type-blind, run:

source ~/.zshrc >/dev/null 2>&1 || true
python3 skills/capacities-lookup/scripts/capacities_cli.py sync-structures

2) Run lookup

Use natural language when helpful.

Examples:

source ~/.zshrc >/dev/null 2>&1 || true
python3 skills/capacities-lookup/scripts/capacities_cli.py lookup "Find my notes on recovery" --json
source ~/.zshrc >/dev/null 2>&1 || true
python3 skills/capacities-lookup/scripts/capacities_cli.py lookup "people associated with an organization" --json

3) Interpret results correctly

Prefer in chat:

  • title
  • type
  • match quality
  • matched-on term when useful
  • clickable markdown deep link

Usually omit raw object ids unless debugging.

Response rules

If a strong match exists

Return the best match directly with a clickable link.

If multiple good matches exist

Return 2–5 candidates.

If the requested object type is not found

Say so clearly. Example:

  • “I didn’t find any Person objects matching that organization, but here are the most relevant fallback Capacities objects.”

If the user asks for relationships/properties/body content

Be explicit about the current API limitation. Example:

  • “I can find the Capacities object and link you to it, but I can’t currently inspect its properties or full body through the public API.”

Operational notes

  • This skill intentionally uses live lookup + light cache, not full sync.
  • Structures metadata is cached under workspace data/capacities/.
  • Lookup cache is convenience-only; live lookup remains authoritative.
  • Type-aware behavior depends on cached structures metadata, so run a structure sync before relying on custom object types.
Usage Guidance
This skill appears to do what it claims: it needs python3 and your Capacities API token and then performs live lookups against the Capacities API, caching lightweight metadata under data/capacities. Before installing or running: (1) verify the CAPACITIES_API_TOKEN is legitimate and only grant it to this tool; (2) inspect config/capacities.json if present (it can set apiBaseUrl and mainSpaceId); (3) avoid blindly running the recommended `source ~/.zshrc` unless you have reviewed that file (sourcing runs shell code, not just setting env vars); (4) note the scripts will write cache/state files under the detected workspace—ensure that location is acceptable. If you want to reduce risk, run the CLI commands manually in a controlled shell rather than granting automatic invocation.
Capability Analysis
Type: OpenClaw Skill Name: capacities-lookup Version: 1.0.0 The capacities-lookup skill is a legitimate tool for searching the Capacities note-taking platform via its public API. The code uses standard Python libraries (urllib) to communicate exclusively with api.capacities.io and implements local caching of metadata in a dedicated data directory. While SKILL.md suggests sourcing shell configuration files (e.g., ~/.zshrc) to retrieve API tokens, this is a common pattern for local agent tools and no evidence of malicious exfiltration, obfuscation, or unauthorized execution was found across the scripts (capacities_client.py, capacities_lookup.py).
Capability Assessment
Purpose & Capability
Name/description match what the code does: the scripts perform live lookups against a Capacities API, build capacities:// deep links, and cache lightweight structure metadata. Declared requirements (python3 and CAPACITIES_API_TOKEN) are appropriate for this purpose.
Instruction Scope
SKILL.md and the CLI instruct the agent/user to source a shell rc file (e.g., source ~/.zshrc) to populate CAPACITIES_API_TOKEN if needed. The scripts themselves only call the Capacities API, read config/capacities.json if present, and write small caches under workspace data/capacities. Sourcing a user shell file can execute arbitrary shell code (not just expose env vars) so follow-up caution is advised before blindly running the provided source commands.
Install Mechanism
There is no external install pipeline or remote download; the skill is distributed as included Python scripts. No URLs/payload downloads or extract steps are used. The scripts will create and write files under the detected workspace data/ directory.
Credentials
Only CAPACITIES_API_TOKEN is required (primary credential). Optionally, CAPACITIES_SPACE_ID or config/capacities.json provides the space id. This is proportional to the stated functionality. Caveat: SKILL.md suggests sourcing ~/.zshrc to load the token which may expose other environment variables or execute rc file logic—review that file before sourcing.
Persistence & Privilege
The skill does not request always:true and will not force-enable itself. It writes caches under the workspace (data/capacities/) and reads/writes its own config/cache/state files only; it does not modify other skills or global agent configuration.
How to Use
  1. Make sure OpenClaw is installed (local or Docker)
  2. Run the install command in chat: /install capacities-lookup
  3. After installation, invoke the skill by name or use /capacities-lookup
  4. Provide required inputs per the skill's parameter spec and get structured output
Version History
v1.0.0
Initial public release
Metadata
Slug capacities-lookup
Version 1.0.0
License MIT-0
All-time Installs 0
Active Installs 0
Total Versions 1
Frequently Asked Questions

What is Capacities Lookup?

Search Capacities for likely object matches and return direct capacities:// links. Use when the user wants to find, open, or locate a Capacities note, meetin... It is an AI Agent Skill for Claude Code / OpenClaw, with 192 downloads so far.

How do I install Capacities Lookup?

Run "/install capacities-lookup" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.

Is Capacities Lookup free?

Yes, Capacities Lookup is completely free, licensed under MIT-0. You can download, install and use it at no cost.

Which platforms does Capacities Lookup support?

Capacities Lookup is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).

Who created Capacities Lookup?

It is built and maintained by Grant Gochnauer (@grantgochnauer); the current version is v1.0.0.

💬 Comments