Append structured markdown entries to a session issues.md log.
npx skills add https://github.com/goodfire-ai/causalab --skill document-issues
Please help me install this Agent Skill. Skill: document-issues Source: https://github.com/goodfire-ai/causalab/tree/main/.claude/skills/document-issues Command: npx skills add https://github.com/goodfire-ai/causalab --skill document-issues
It prevents repeated pain by turning failures, confusions, odd results, and workarounds encountered during work into a persistent, session-level record that future agents can learn from.
Run /document-issues while you have an active research session so the skill updates agent_logs/<SESSION_NAME>/issues.md with what went wrong and how you worked around it.
High-intent search queries and answers about installing and using this skill.
To log debugging issues and workarounds, run the command during an active research session. It appends structured markdown entries to the session's issues.md file, capturing failures, oddities, and fixes for future agents to reference.
A persistent record of environment problems is maintained by writing structured markdown logs into an active session directory. This approach proactively captures errors and unexpected behavior, preventing repeated debugging pain across future agent runs.
You document unexpected behavior by appending categorized markdown entries to issues.md. The skill resolves the correct session directory using an active marker and records the failures, confusions, and workarounds encountered during your workflow.
Yes, you need a valid agent session marker located at agent_logs/.current. The skill refuses to run if this marker is missing or stale, prompting you to start a research session first to ensure proper logging.
The issues log command fails because it requires a valid active session marker at agent_logs/.current to resolve the correct directory. Without it, the skill cannot determine where to append the structured markdown entries for your workarounds.
You should not use a persistent issues log if you lack an active research session. The operational guardrails prevent execution when the agent_logs/.current marker is missing, ensuring entries are never written to incorrect or stale directories.
goodfire-aiCreate a session directory with plan, run, and result scaffolding for research workflows.
Ricardo-VaeTrace research code flow and validate edge cases and numerical correctness.
mentileadCapture and categorize agent actions, escalations and failures into dated markdown files.
e-stpierreAppend structured NDJSON log entries with ISO 8601 timestamps to workflow logs.
