Strip inline <private> blocks and classify is_private flags before storage.
npx skills add https://github.com/paulingham/.claude --skill capture-paulingham
Please help me install this Agent Skill. Skill: capture Source: https://github.com/paulingham/.claude/tree/main/skills/capture Command: npx skills add https://github.com/paulingham/.claude --skill capture-paulingham
Capture-time privacy is achieved by stripping <private> blocks and applying an allowlist classification before data is inserted into observations or scratchpad_findings, using only the Python standard library.
Apply the capture privacy workflow to a new data envelope to scrub private blocks and classify privacy before storage.
High-intent search queries and answers about installing and using this skill.
Capture-time sanitization removes inline <private> tags and their contents from text fields before data reaches disk. This ensures sensitive content is stripped during ingestion, preventing unauthorized exposure in stored observations.
Allowlist-based classification assigns is_private flags by matching sanitized file paths against globs or content against regex patterns. This gates recall-time visibility, ensuring only authorized data is accessible during retrieval.
Yes, privacy scrubbing operates using Python stdlib only, requiring no external dependencies. It integrates with capture write surfaces to strip <private> blocks and apply allowlist classification before storage.
Sanitization occurs before content_hash computation to guarantee deterministic storage integrity. This ensures identical hashes for the same envelope across both live and replay paths, preventing hash mismatches.
Apply an allowlist during ingestion to set is_private flags based on sanitized file path globs or content regex matches. This gates recall-time visibility, restricting access to sensitive observations.
Yes, the capture privacy workflow integrates with write surfaces for both observations and scratchpad_findings. It scrubs <private> blocks and applies allowlist classification before data is inserted into either storage target.
angryracoonPrepare Salesforce Data Cloud ingestion assets for raw source data.
rubenslivQuery CZ CELLxGENE Census single-cell data via TileDB-SOMA in Python.
nimadorostkarOrganize and deduplicate file systems using content-based hashing with reversible relocation.
ppiankovDetect and obfuscate sensitive data in text using deterministic regex methods.
