From pluggedin
Record completion of a step from a written implementation plan into fresh memory. Use after marking a plan item done — maintains task continuity across context compaction.
How this skill is triggered — by the user, by Claude, or both
Slash command
/pluggedin:memory-capture-plan-stepThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
> **SECURITY — Two-Layer Defence (PCI-DSS 4.0 · ISO/IEC 27001:2022 · GDPR · SOC 2)**
SECURITY — Two-Layer Defence (PCI-DSS 4.0 · ISO/IEC 27001:2022 · GDPR · SOC 2)
Layer 1 (automatic): The hook script pipes all content through
pci-scrub.pybefore sending to the API. Catches: emails, API keys (OpenAI/Anthropic/AWS/GitHub/…), IPv4/IPv6, credit card numbers, national IDs (US SSN, UK NI, EU VAT, Aadhaar, …), connection string credentials, home paths, phone numbers.Layer 2 (YOU): Regex cannot catch everything. Before calling any memory tool, apply semantic judgment. Do NOT record:
- Person names, usernames, employee IDs → replace with
[PERSON]- Company / customer names →
[COMPANY]- Passwords, MFA codes, session tokens, cookies →
[REDACTED]- Health / medical data (HIPAA PHI) →
[REDACTED]- Financial account numbers, balances →
[REDACTED]- Precise location data (GPS, home address) →
[ADDRESS]- Anything business-confidential (pricing, contracts) →
[REDACTED]When in doubt, leave it out. Over-redaction is always preferable to a compliance violation.
Record plan step completion to maintain continuity across context resets.
PLAN: <plan name / file>
STEP: <step number and title>
STATUS: completed
WHAT WAS DONE:
<concise description of exactly what was implemented>
DEVIATIONS FROM PLAN:
<any differences from the original plan spec, or "none">
NEXT STEP:
<what comes next in the plan>
BLOCKERS/NOTES:
<anything the next session needs to know, or "none">
pluggedin_memory_observe with:
sessionUuid: current session UUIDobservationType: workflow_stepcontent: the structured textoutcome: "success"metadata: {"ring": "fresh", "plan_step": "<step number>", "plan_name": "<name>"}After context compaction, running memory-resume will surface recent plan steps.
This is what lets a new session know "Step 3 is done, Step 4 is next" without
re-reading the entire plan from scratch.
PLAN: memory-system-overhaul (2026-03-09)
STEP: Phase 1, Week 1 — PCI Scrubber
STATUS: completed
WHAT WAS DONE:
Created pci-scrub.py in pluggedin-plugin/plugin/scripts/.
Covers: emails, API keys, credit cards, tax IDs, IPs, home paths, phones.
Updated observe-tool-result.sh to pipe through scrubber before sending to API.
DEVIATIONS FROM PLAN:
None — implemented exactly as designed in Design Section 1.
NEXT STEP:
Phase 1, Week 1 — 6 Trigger Skills (memory-capture-solution, etc.)
BLOCKERS/NOTES:
none
Skip this skill if:
ls, checked a file, viewed a README)<private> block wraps the relevant content — respect user opt-outnpx claudepluginhub veriteknik/pluggedin-plugin --plugin pluggedinSaves facts, decisions, and context to Memsy memory for future sessions. Strips framing verbs and persists the substance verbatim via memsy_ingest.
Persists task plans, findings, and progress as markdown files so work survives context loss and /clear. Use for multi-step projects or tasks requiring 5+ tool calls.
Captures durable knowledge from sessions into rekal memory with deduplication. Extracts preferences, decisions, architecture, and procedures; checks for duplicates; stores or supersedes.