Search past agent conversations to recall prior discussions, decisions, and context. Activate this skill when users ask about previous conversations, want to resume past work, or reference earlier decisions.
Instal·la
npx skillscat add allenhutchison/obsidian-gemini/recall-sessions Instal·la des del registre de SkillsCat.
Recall Past Sessions
Find and retrieve past agent conversations using the recall_sessions tool. This enables recalling prior discussions, decisions, and context from earlier sessions.
When to Use
Use this skill when the user:
- Asks "what did we discuss about..." or "last time we talked about..."
- Wants to resume or continue previous work
- References a decision or conversation from a past session
- Asks "have we worked on..." or "do you remember when..."
- Needs context from a prior session about a specific file or project
How to Use
Call the recall_sessions tool with these parameters:
query(optional) — Search term to match against session titles (case-insensitive substring match).filePath(optional) — Find sessions that accessed a specific file (e.g.,notes/meeting.md).project(optional) — Find sessions linked to a specific project name.limit(optional) — Maximum results to return. Default is 10, maximum is 50.
At least one of query, filePath, or project should be provided for meaningful results.
Search Strategies
- Topic search — Use
querywhen the user asks about a discussion topic:recall_sessions(query="refactoring") - File-based recall — Use
filePathwhen the user references a specific file:recall_sessions(filePath="projects/website-redesign.md") - Project recall — Use
projectwhen the user mentions a project:recall_sessions(project="website-redesign") - Combined search — Use multiple parameters to narrow results:
recall_sessions(query="API design", project="backend")
Progressive Disclosure
The tool returns session summaries only — title, date, files accessed, project linkage, and a historyPath.
To see the full conversation from a past session, use read_file on the returned historyPath. This two-step approach avoids loading unnecessary conversation data.
Workflow
- Call
recall_sessionsto find relevant sessions - Review the returned summaries with the user
- Use
read_fileon thehistoryPathof the session(s) the user is interested in - Summarize or reference the relevant parts of the conversation
Tips
- Results are sorted by most recent first — the newest sessions appear at the top.
- The current session is automatically excluded from results.
- When the user asks a vague question like "what did we do last time", try a broad
queryor search by the files currently in context usingfilePath. - If no results are found, let the user know and suggest alternative search terms or parameters.