Menu
estimate CC active time for a task using the time rule, with dynamic effort resolution
Install with Codex or Claude Copy this prompt, paste it into Codex, Claude, or another assistant, and let it review the skill page and install it for you.
Based on SOC occupation classification
See active Claude Code sessions on this machine and coordinate across them via direct messages, status announcements, and file-overlap awareness. Triggers when the user asks "who else is running?", "any other claude open?", "what other sessions are active?", "tell <name> to ...", "ping <peer>", "let peers know I'm doing X", "broadcast <status>", "check the cc digest", or any cross-session / multi-agent coordination on this machine.
measure your real CC throughput against the time rule's matrix, using lore.db
export the lore knowledge graph or its slices to json, csv, or markdown
explore relationships in the lore knowledge graph - cooccurring files, project sessions, sibling sessions, tagged notes
query the lore knowledge graph for sessions, costs, tools, projects, search, and patterns
capture decisions, lessons, reminders, and todos into the lore knowledge graph
| name | time-estimate |
| description | estimate CC active time for a task using the time rule, with dynamic effort resolution |
| argument-hint | <task description> |
| user-invocable | false |
produce a realistic CC active-time estimate for the given task, grounded in plugins/cc/rules/time.md. never guesses effort level.
resolve active effort, highest-precedence first. stop at the first rung that returns a value:
echo "${CLAUDE_CODE_EFFORT_LEVEL:-}" (env var)--effort flag, if visible via process context/effort session state, if already set this sessioneffortLevel key in ~/.claude/settings.json, then any project-scoped .claude/settings.json under the repo rootcite the rung that supplied the answer. if ambiguous, ask the user rather than guess.
resolve active model from the session's model id (e.g. claude-opus-4-7).
classify session mode using the table in the rule:
apply the multiplier from the model × effort matrix in the rule. quote the baseline at low, then scale.
return the estimate in the rule's format:
CC: <range> (always a range, never a single number)if the task breaks into sub-phases, estimate each and sum. flag which phases parallelize via subagents (1.8-2.2×) and which stay serial.
xhigh or low without checking.xhigh is ~30% slower than high for throughput-bound work.task: add logout button with tests
effort: low (rung 4, ~/.claude/settings.json)
model: claude-opus-4-7 (0.87× baseline)
mode: standard
CC: 12-18 min active
your time: ~10 min review
confidence: medium
phases:
P1 find auth module (serial): 2-3 min
P2 write button + route (serial): 4-6 min
P3 tests (parallel via subagent): 3-5 min wall, ~6-9 min work
P4 verify (serial): 3-4 min
risks that 2×: auth module structure unclear, test framework unfamiliar
the rule's matrix is a generic baseline. for personal calibration, run /time-calibrate.