Skip to content

REMINDERS.md

Latest commit

 

History

History
85 lines (60 loc) · 2.14 KB

REMINDERS.md

File metadata and controls

85 lines (60 loc) · 2.14 KB

clog Contextual Reminders

This document shows what reminders appear for different event types and states.

Task Events

Task Started (with user-prompt)

$ clog -type=task -state=in_progress -user-prompt="Fix the login bug" -message="Debugging auth flow" -session="bugfix"
200 OK

  Remember: Log completion when done: clog -type=task -state=completed -message="..." -session="..."

Task Started (without user-prompt)

$ clog -type=task -state=in_progress -message="Debugging auth flow" -session="bugfix"
200 OK

  TIP: Add -user-prompt="<exact user text>" to capture full context (VERBATIM user input)

Task Completed

$ clog -type=task -state=completed -message="Login bug fixed" -session="bugfix"
200 OK

  Next: Log next task or use -type=progress for multi-step work

Task Blocked

$ clog -type=task -state=blocked -message="Waiting for API key" -session="bugfix"
200 OK

  Consider: Use -type=question -state=blocked if waiting for user input

Question Events

Question (blocked)

$ clog -type=question -state=blocked -message="Should I use OAuth2 or SAML?" -session="bugfix"
200 OK

  Good: Always log questions before asking user

Progress Events

Progress Update

$ clog -type=progress -message="50% complete (5/10 tasks)" -session="bugfix"
200 OK

(No contextual reminder - progress updates are informational)

Session Events

Session Started

$ clog -type=session -message="Starting API improvements" -session="api-work"
200 OK

  Remember: Log tasks with -type=task as you work

Session Completed

$ clog -type=session -state=completed -message="API improvements complete" -session="api-work"
200 OK

(No contextual reminder - session is ending)

Design Rationale

Character economy: Every character costs money in LLM context windows.

Contextual: Reminders only appear when relevant to the current event type/state.

Actionable: Each reminder shows concrete next steps.

Positive reinforcement: "Good:" prefix confirms correct behavior (e.g., logging questions).

No emojis: Removed to save characters without losing clarity.