| name | tools |
| description | Show which tools and commands you rely on most across coding sessions — tool histogram, top shell commands, and Read:Write ratio. Frequency, not duration (that's time) or cost (that's cost). Use when the user asks "what tools do I use most", "my top commands", "what do I reach for", "how often do I edit vs read", or "tool breakdown". |
tools
What you reach for — the frequency of tools and commands across your sessions.
Counts, not time (/openstory:time) or money (/openstory:cost). Two phases:
pull via MCP, then render. Report only what the tools return.
Phase 1 — pull the numbers
Prefer a dedicated histogram tool if it exists:
mcp__openstory__tool_histogram — aggregate tool + command counts over a
window. May not exist yet (it would wrap /api/insights/tool-evolution
plus bash-command parsing). If present, use it.
Fallback (existing tools, mcp__openstory__*):
mcp__openstory__list_sessions — pick the heaviest recent sessions.
mcp__openstory__tool_journey on those — per-session tool sequence; aggregate
the counts across them into a tool histogram. Tool-level only — command-level
(grep/git/ssh) needs the dedicated tool above.
Default to the last 7 days.
If those tools aren't available, OpenStory isn't connected — say so, don't guess.
Phase 2 — render
What you reach for — last <N> days
Read:Write <ratio> · most-used: <tool> (<pct>%)
Tools
<tool> <count> <bar>
...
Top commands # only with tool_histogram (bash parsing)
<cmd> · <count> ...
Close with one line: your single most-used tool, and the Read:Write balance
(read-heavy = careful; write-heavy = fast iteration).
Honest limit
Without tool_histogram, this is tool-level only (Bash/Read/Edit), composed
from a sample of heavy sessions — command-level frequency needs the server tool.
Say which mode you ran in.
When NOT to use this skill
- For time spent (duration), use
/openstory:time.
- For token/dollar cost, use
/openstory:cost.