Sequential Read
PassAudited by ClawScan on May 1, 2026.
Overview
Sequential Read appears to be a coherent local reading tool, with the main thing to notice being that it stores reading data and runs follow-up sub-agents automatically after you start a session.
This skill looks reasonable for its stated purpose. Before installing or using it, be aware that it can run a hands-off multi-agent reading session and persist copies or derivatives of the selected text under memory/sequential_read; avoid sensitive files unless that local storage is acceptable.
Findings (3)
Artifact-based informational review of SKILL.md, metadata, install specs, static scan signals, and capability signals. ClawScan does not execute the skill or run runtime probes.
The skill will process the file path you provide and create local outputs from it.
The skill runs local Python helper scripts with the user-provided source path and optional lens. This is expected for the reading workflow, but it means the session has authority to read that chosen local file and write derived session data.
python3 {baseDir}/scripts/session_manager.py create <source-file> [--lens <persona>]Invoke it only on files you intend to analyze, and avoid using it on sensitive documents unless you are comfortable with local session copies being created.
Private or copyrighted text submitted to the skill may remain in the local OpenClaw workspace along with reflections and synthesis output.
Chunk text is saved into a persistent workspace memory directory for the session, so source material and derived reflections can remain available after the run.
return get_workspace() / "memory" / "sequential_read" / session_id ... chunk_path(sd, n).write_text(text, encoding="utf-8")
Use it with documents you are comfortable storing locally, and delete the relevant memory/sequential_read session directory when you no longer need it.
A long reading task may continue autonomously and may consume additional model time while sub-agents read and synthesize the selected file.
The skill explicitly uses spawned reader and finisher sub-agents and continues without additional confirmation once a reading session starts. This is disclosed and scoped to completing the reading pipeline.
The pipeline runs in spawned sub-agents... Do not wait or ask the user between the main reader and finisher... The whole pipeline should be hands-off.
Start sessions only when you want the full pipeline to run, and monitor or cancel long jobs if cost, time, or document sensitivity matters.