analyzing-task-runs — Claude Code skill: run analysis
analyzing-task-runs — a Claude Code skill that breaks down a completed task run from its log and records the activities. It starts when asked to analyze a run or review an attached log.
no license 148 installs ★ 101 updated 2026-10-24
npx skills add PostHog/ai-plugin --skill analyzing-task-runs -a claude-codeWhat analyzing-task-runs does
The skill divides a task log into activities, capturing the agent’s goal, outcome, any blocker, and supporting lines. It submits each record through report_activity, then lists the recorded activities and calls finish with status completed.
It needs an attached .jsonl run log under .posthog/attachments. The skill identifies the log format and uses bounded jq queries; it handles pi and ACP schemas. The log must be queried in capped sections rather than read unfiltered. Records contain facts, not repair suggestions.
What it helps with
- split an agent run into goal-based activities
- record each activity’s outcome and blocker
- attach log lines as evidence for an activity
- mark where a user redirect changes the task
Example prompts for the agent
- Analyze the completed run from the attached log and record its activities.
- Review this run log, noting the agent’s actions and any blockers.
- Break down the attached
.jsonltask log into activities.
How to install analyzing-task-runs
1Skills CLI (Vercel)
npx skills add PostHog/ai-plugin --skill analyzing-task-runs -a claude-code2Manually: put the files in place
~/.claude/skills/analyzing-task-runs/SKILL.mdCopy the whole skill folder; for one project use .claude/skills/analyzing-task-runs/
1Skills CLI (Vercel)
npx skills add PostHog/ai-plugin --skill analyzing-task-runs -a codex2Manually: put the files in place
.agents/skills/analyzing-task-runs/SKILL.mdFor all projects use ~/.agents/skills/
1Skills CLI (Vercel)
npx skills add PostHog/ai-plugin --skill analyzing-task-runs -a cursor2Manually: put the files in place
.cursor/skills/analyzing-task-runs/SKILL.mdCursor also reads .agents/skills/ and .claude/skills/
1GitHub CLI
gh skill install PostHog/ai-plugin skills/analyzing-task-runs2Manually: put the files in place
.github/skills/analyzing-task-runs/SKILL.md1Skills CLI (Vercel)
npx skills add PostHog/ai-plugin --skill analyzing-task-runs -a gemini-cli2Platform CLI
gemini skills install https://github.com/PostHog/ai-plugin --path skills/analyzing-task-runsManually: put the files in place
.devin/skills/analyzing-task-runs/SKILL.mdThe legacy .windsurf/skills/ path still works
Skills CLI (Vercel)
npx skills add PostHog/ai-plugin --skill analyzing-task-runs -a opencodeClaude.ai
Zip the folder and upload it in Settings → Capabilities → Skills
Step by step, with commands for each platform: How to install skills in Claude Code: five ways
Questions about analyzing-task-runs
How do I install analyzing-task-runs in Claude Code?
Run npx skills add PostHog/ai-plugin --skill analyzing-task-runs -a claude-code. The skill file is at ~/.claude/skills/analyzing-task-runs/SKILL.md.
What does analyzing-task-runs do?
It turns a completed run log into activity records and submits them through report_activity. It calls finish when the analysis is done.
What does analyzing-task-runs need to work?
It needs an attached .jsonl run log and access to the report_activity and finish tools. Supported log schemas are pi and ACP.