Routines Catalog

Routines are periodic self-maintenance tasks that agents check at session start. When a routine is overdue, the agent flags it — but execution is always opt-in. The user decides whether to run it now, defer, or skip.

All Routines

Name	Description	Frequency	Compatible Agents
backlog-hygiene	Review and maintain Playbook/Backlog.md — flag stale items, escalate P0s, cross-reference with Status and Roadmap	7 days	tech-lead
dependency-audit	Scan dependencies for known vulnerabilities and unmaintained packages	7 days	tech-lead
doc-freshness-check	Detect documentation drift — docs that are stale relative to recent code changes	7 days	tech-lead
infra-drift-check	Detect infrastructure drift — compare declared IaC state against actual cloud resources	7 days	tech-lead
memory-consolidation	Bridge Claude Code auto-memory into agent/Core/memory.md — validate facts, clean redundant entries	5 days	tech-lead
roadmap-accuracy	Ensure Playbook/Roadmap.md reflects what’s actually built and what’s actually next	14 days	tech-lead
status-hygiene	Keep Playbook/Status.md clean — archive old Done items, validate Pending items	5 days	tech-lead
vulnerability-scan	SAST scan, secrets scan, and dependency audit — flag new, resolved, and persistent vulnerabilities	7 days	tech-lead

Routine Details

backlog-hygiene

Reviews Playbook/Backlog.md for stale items that have been sitting too long, escalates any P0 items that need immediate attention, and cross-references against Status.md and Roadmap.md to ensure consistency. Runs every 7 days.

dependency-audit

Scans the project’s dependency files (go.mod, package.json, etc.) for known vulnerabilities using available audit tools and flags unmaintained packages. Runs every 7 days.

doc-freshness-check

Compares documentation files against recent code changes to detect drift. Flags docs that reference files, functions, or structures that have changed since the doc was last updated. Runs every 7 days.

infra-drift-check

Compares declared infrastructure-as-code state (Terraform plans, Kubernetes manifests) against actual cloud resources to detect configuration drift. Runs every 7 days.

memory-consolidation

Reviews Claude Code’s auto-memory entries and validates them against the codebase. Migrates accurate facts into the agent’s version-controlled agent/Core/memory.md, cleans redundant entries, and archives outdated information. Runs every 5 days.

roadmap-accuracy

Checks Playbook/Roadmap.md against the current state of the codebase and completed work. Ensures the roadmap reflects what has actually been built and what is genuinely planned next. Runs every 14 days.

status-hygiene

Cleans up Playbook/Status.md by archiving old “Done” items, validating that “Pending” and “In Progress” items are still accurate, and removing stale entries. Runs every 5 days.

vulnerability-scan

Performs a comprehensive security scan: static analysis (SAST), secrets detection, and dependency vulnerability audit. Flags new vulnerabilities, tracks resolved ones, and monitors persistent issues. Runs every 7 days.

How Routines Work

Session start — The routine-check sensor parses the routine dashboard (agent/Core/routines.md), compares last_ran dates against each routine’s frequency, and flags any that are overdue.
User decides — The agent presents overdue routines. The user chooses to run now, defer, or skip.
Execution — The agent reads the routine’s definition file in agent/Routines/ and follows the procedure step by step.
Log — Results are appended to Logs/RoutineLog.md with the date, routine name, outcome, and notes.
Update — The last_ran date in the dashboard is set to today.

Default Routines by Agent Type

Agent Type	Default Routines
Tech Lead	(none by default — all 8 are compatible and selectable)
DevOps	dependency-audit, infra-drift-check
Security	vulnerability-scan, dependency-audit
Backend	(none)
Frontend	(none)
Full-Stack	(none)