memory-capture-plan-step

name	memory-capture-plan-step
description	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.
user-invocable	true

Memory Capture — Plan Step

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.py before 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.

When to Use

• A numbered item in a written implementation plan was just completed
• A phase / week milestone was reached
• You are about to mark a checklist item as done

What to Capture

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">

How to Record

1. Build the observation text above
2. Scrub PII before recording
3. Call pluggedin_memory_observe with:
   • sessionUuid: current session UUID
   • observationType: workflow_step
   • content: the structured text
   • outcome: "success"
   • metadata: {"ring": "fresh", "plan_step": "<step number>", "plan_name": "<name>"}

Why This Matters

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.

Example

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 Guidance

Skip this skill if:

• The information is already obvious from the current conversation context
• The content contains personal names, customer data, or business-confidential information that cannot be fully redacted
• The observation is trivial (e.g., ran ls, checked a file, viewed a README)
• A <private> block wraps the relevant content — respect user opt-out