Promote high-value observations from daily logs to long-term memory.
npx skills add https://github.com/adiman9/mnemos --skill consolidate-adiman9
Please help me install this Agent Skill. Skill: consolidate Source: https://github.com/adiman9/mnemos/tree/main/core/skills/consolidate Command: npx skills add https://github.com/adiman9/mnemos --skill consolidate-adiman9
Promote high-value working memory observations to the long-term knowledge pipeline, ensuring important insights persist across sessions and deployments.
Run consolidate on the latest daily logs to promote eligible observations to long-term memory.
High-intent search queries and answers about installing and using this skill.
To promote working memory observations to long-term knowledge, this Skill scans recent daily logs, applies promotion criteria, and batches high-value items through the reduce pipeline. This ensures important insights persist across sessions and deployments.
Consolidating AI agent memory involves scanning daily logs in your vault to identify high-value observations, applying promotion criteria, and routing promoted data through identity, reference, and full-pipeline paths to enrich the knowledge graph.
You can preview memory promotions before committing changes by using the dry-run flag. This allows you to scan the default last 7 days of daily logs and see which observations meet the promotion criteria without writing to long-term memory.
Yes, you can specify a custom time range for scanning daily memory logs by using the days flag. The Skill defaults to scanning the last 7 days of logs in the vault_path/memory/daily directory, but this can be adjusted.
Automating long-term memory management requires vault metadata configured in a .mnemos.yaml file. The Skill reads this configuration to locate daily logs and writes promotion results to memory, notes, self, and inbox paths.
Automated memory consolidation relies on specific vault structures in vault_path/memory/daily and .mnemos.yaml metadata. It is limited to processing observations from daily logs and does not extract high-value insights from unstructured or non-standardized input formats.
adiman9Extract typed observations from session transcripts and score them by importance, confidence, and surprise.
solpbcManage tracked people, companies, and projects via command-line operations.
just-cameronObserve ATProtocol firehose activity and record engagement patterns as cognition thoughts.
roberttmadsen13-delExtract and store project or conversation insights as searchable memories in MemPalace.
