Claude Skill MartBrowse skillsQuick linesImage promptsLearn by videoTerminal guideWhat is a Skill?
 
← Back to list

CASS Session Log Search

Operating doctrine for using the `cass` CLI to search past agent session logs and return cited prior prompts, decisions, and failures.

UtilitiesAdvanced★ 445⑂ 41AI score 8/10Last updated: Sep 24, 2026

What it does

  • Searches your archived AI agent conversation logs through the cass CLI and returns evidence with source path + line number.
  • Enforces a fixed discovery workflow: search → pack (cited excerpts) → view/expand (verify the raw record) → context (find related sessions).
  • Distinguishes three index states — healthy, stale, missing — so a merely stale index is searched immediately instead of triggering a wasteful full rebuild.
  • Encodes interpretation rules: zero hits means "no match within these query limits", not global absence; a repeated prompt is not proof it worked; failure sessions outweigh success sessions when mining prior art.
  • Ships an anti-pattern table from real incidents: never run bare cass (blocking TUI), never --limit 0, never pipe cass export into head, don't treat subagent logs as parent sessions.

Who it's for

  • Developers with a long history of agent harness sessions (Claude Code and similar) worth mining.
  • Anyone who regularly asks "what did I ask last time?" or "when did we decide not to do X?".
  • Users recovering working context after a crash or context loss.
  • CLI-comfortable users willing to install the external cass binary and jq.

Examples

  1. Prior-art check: before adopting an approach, run cass search "retry backoff" --mode lexical --json --fields minimal --limit 20, deliberately hunting for failure strings like "reverted" or "didn't work" first.
  2. Decision archaeology: search for when a scope decision was made, then open the exact record with cass view PATH -n LINE -C 5 to confirm speaker role and actual outcome.
  3. Post-crash recovery: run timeout 15 cass status --json, confirm the index is usable, and re-find the keywords you were chasing before the session died.

Note: this skill is doctrine, not the tool. It requires the external cass binary (>= 0.3.6 recommended) and jq; semantic mode additionally needs a ~90MB model bundle.

· · · Install guide · · ·

Try it now, no install

Paste this into Claude to use the skill without installing anything.

Read the instructions in this file and follow them to help me:
https://raw.githubusercontent.com/boshu2/agentops/HEAD/images/gemini/skills/cass/SKILL.md

What I want: (describe your task here)

If Claude can't open the link, open it yourself and paste the contents instead.

↓ If it works for you, download the ZIP below and install it. Then it runs on its own — no pasting each time.

Install in the Claude app (no terminal)
  1. Download the ZIP with the button below.
  2. In Claude, open Settings → Capabilities and turn on 'Code execution and file creation'. (one time)
  3. Go to Customize → Skills → + → 'Upload a skill' and upload the ZIP.
↓ Download ZIP
Install in Claude Code

Let Claude do it — paste this into Claude Code

Install the skill I found on Claude Skill Mart.
Copy the images/gemini/skills/cass folder from the GitHub repo boshu2/agentops into my ~/.claude/skills/cass/.
When it's done, tell me in one line what this skill can do.

Install with a command instead

git clone https://github.com/boshu2/agentops.git && mkdir -p ~/.claude/skills && cp -r agentops/images/gemini/skills/cass ~/.claude/skills/cass

⚠ This is a third-party skill. Check the source repository before installing.

  1. Open a terminal.
  2. Install the prerequisites first: the upstream cass binary (Dicklesworthstone project) — verify with cass --version — and jq for JSON parsing (brew install jq on macOS, sudo apt install jq on Ubuntu).
  3. Clone the repository: git clone https://github.com/boshu2/agentops.git
  4. Create the skills directory: mkdir -p ~/.claude/skills
  5. Copy the skill in: cp -r agentops/images/gemini/skills/cass ~/.claude/skills/cass
  6. Confirm the references/ and scripts/ subfolders came along — the skill body links to them heavily.
  7. Restart Claude Code and trigger it with a request like "search my past sessions for what I asked about X".
  8. Validate the setup by running ~/.claude/skills/cass/scripts/validate.sh.
  9. Warning: never launch bare cass inside an agent loop — it opens a blocking TUI. Always pass --json.
View source on GitHub ↗License: Apache-2.0