Stats
Actions
Tags
From forge-memory
Tiered memory lookup strategy for decoding workplace shorthand, acronyms, and internal language. Transform "ask todd about PSR" into full context using taxonomy and knowledge files.
How this skill is triggered — by the user, by Claude, or both
Slash command
/forge-memory:memory-managementThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
Memory makes Claude your workplace collaborator—someone who speaks your internal language and decodes shorthand like a colleague would.
Memory makes Claude your workplace collaborator—someone who speaks your internal language and decodes shorthand like a colleague would.
Transform shorthand into understanding:
User: "ask todd to do the PSR for oracle"
↓ Claude decodes
"Ask Todd Martinez (Finance lead) to prepare the Pipeline Status Report
for the Oracle Systems deal ($2.3M, closing Q2)"
Without memory, that request is meaningless. With memory, Claude knows:
Always decode shorthand before acting on requests using a tiered approach:
Query organizational taxonomy for:
Use forge memory get-taxonomy to query these.
Coverage: Official organizational entities (products, clients, teams, integrations).
Search memory/glossary.md for:
Lifecycle filtering: Entries have an importance score (0-100) that decays over time.
PSR → Pipeline Status Report (fading)Coverage: All workplace shorthand and decoder ring entries.
Search rich detail files:
memory/people/{name}.md — Full person profilesmemory/projects/{name}.md — Project detailsmemory/context/ — Company, tools, processesLifecycle filtering: Same rules as Tier 2 apply to knowledge files with frontmatter importance scores. Skip sunset entries, flag probationary ones with (fading).
Coverage: Execution-level context and detailed profiles.
If not found in any tier:
I don't know what X means yet. Can you tell me?
I'll remember it for next time.
Note: Before asking the user, check memory/pending.json — the term may already be tracked there from a previous mention but not yet promoted to knowledge. If found in pending, mention it: "I've seen X mentioned before but don't have a definition yet. Can you clarify?"
User: "ask todd about the PSR for phoenix"
1. Tier 1 (Taxonomy via forge-lib):
→ Products? Teams? Clients? (no match)
2. Tier 2 (Glossary):
→ "todd" → Todd Martinez, Finance lead ✓
→ "PSR" → Pipeline Status Report ✓ (importance 55, trusted)
→ "phoenix" → (not in glossary)
3. Tier 3 (Deep Memory):
→ Search memory/projects/ for "phoenix"
→ Found: memory/projects/phoenix.md (importance 25, probationary)
→ Phoenix = DB migration project ✓ (fading)
4. Boost recalled entries:
→ todd, PSR, phoenix all get +5 importance
Now Claude can act with full context.
Note: "phoenix" flagged as (fading) — consider confirming it's still active.
Source: forge memory get-taxonomy queries
Update: /memory:setup-org command
Source: Direct file reads from memory/ directory
Update: /memory:remember and /memory:recall commands
Start with fast lookups (taxonomy), expand to detailed files only when needed.
When presenting a result, note related entries:
Combine information from multiple tiers to build complete understanding:
"todd" from glossary + memory/people/todd-martinez.md
= Todd Martinez, Finance lead, prefers Slack, handles PSRs
When encountering unknown terms:
Successful recalls automatically strengthen entries — each recall reinforces importance, keeping frequently-used knowledge in trusted status:
Taxonomy (via forge-lib):
Glossary entries:
Deep memory:
Preferences:
Fast Tier (Taxonomy + Glossary):
Detail Tier (Deep Memory):
This tiered strategy keeps common lookups fast while supporting unlimited depth.
All file operations delegated to forge-lib or direct file reads—this skill provides the reasoning strategy only.
npx claudepluginhub jeremybrice/the-forge --plugin forge-memoryImplements two-tier memory with CLAUDE.md hot cache and memory/ directory to decode acronyms, nicknames, people, projects, and internal terms.
Activate for: workplace memory, person brief, team profile, organisation profile, add person, add project, add terminology, four-layer memory, communication style, stakeholder brief, post-meeting memory revision, who works on, team directory, revise workplace memory, organisation context. NOT for: basic memory CRUD (use official productivity plugin memory-management), cross-context search (use workplace-search).
Searches Confluence, Jira, and internal docs for company systems, terminology, processes, deployment, authentication, infrastructure, and technical concepts with parallel cited results.