Researcher Context Audit
ProductivityResearcher Context Audit is an Agent Skill for productivity published in the SpectrAI-Initiative/InnoClaw repository on GitHub: a SKILL.md instruction file an AI agent can load to do this task.
Use when the Researcher starts, resumes, or reaches a major decision point. Build a context inventory from workstation materials, prior messages, existing artifacts, requirements, and unfinished work.
How to use this skill
Bring this guide into your coding agent with a prompt tailored to the tool you use.
- Open your project in Codex.
- Copy and paste the prompt below.
- Review files and risks before approving any action.
Inspect the Agent Skill at https://github.com/SpectrAI-Initiative/InnoClaw/blob/HEAD/.claude/skills/researcher-context-audit/SKILL.md and all supporting files. Report provenance, permissions, scripts, dependencies, and security risks. Propose an exact installation into .agents/skills/researcher-context-audit, but wait for my approval before writing files or running scripts.
Copying this prompt does not install or run the skill. Review third-party files before use.
Researcher Context Audit
Use this skill before planning, before major re-planning, and after interruptions. The goal is to establish a trustworthy operating picture for the automated research workflow.
What To Review
- User goal, preferred framing, and explicit constraints.
- Existing nodes, artifacts, and checkpoints.
- Completed, pending, blocked, superseded, and rejected work.
- Research memory, workstation materials, recent notes, and prior decisions.
- Evidence quality, open questions, and unresolved contradictions.
Output Contract
Produce a compact but decision-ready context review that includes:
- Confirmed objective.
- What has already been completed.
- What remains unfinished.
- Which assumptions are still unconfirmed.
- Which prior directions were rejected or superseded.
- Which constraints or resources materially affect the next step.
Quality Rules
- Distinguish facts from summaries and summaries from inferences.
- Do not repeat the full transcript; compress to the minimum useful operating state.
- Do not ignore prior user corrections or prior failed attempts.
- If the context is insufficient for safe planning, explicitly trigger the ambiguity gate instead of improvising.