Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add OutlineDriven/outline-driven-development --skill continual-learninggit clone --depth 1 https://github.com/OutlineDriven/outline-driven-developmentWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/outlinedriven/outline-driven-development/continual-learning)<a href="https://agentmods.dev/skills/outlinedriven/outline-driven-development/continual-learning"><img src="https://agentmods.dev/badge/skills/outlinedriven/outline-driven-development/continual-learning/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/outlinedriven/outline-driven-development/continual-learning"><img src="https://agentmods.dev/badge/skills/outlinedriven/outline-driven-development/continual-learning.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector pass
What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00036 | $0.00839 |
| Opus 5 | $0.00018 | $0.00419 |
| Sonnet 5 | $0.00007 | $0.00168 |
| Haiku 4.5 | $0.00004 | $0.00084 |
Grade A, and why
continual-learning scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 2d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 46 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Continual learning
Contract
| Field | Bound contract |
|---|---|
| Trigger | A scheduled tick or watcher event fires to mine prior chats and maintain project memory. |
| Authority | Reversible local: writes only AGENTS.md and the continual-learning index; rollback is version control. No remote mutation. |
| Side effect | Updates AGENTS.md and the continual-learning index with deduplicated high-signal memory entries. |
| Done | Deduplicated high-signal memory updates are written, or an explicit no-update result is returned. |
Inputs
- Prior chat transcripts or session logs accessible in the local workspace (required).
- Existing AGENTS.md (required, read before mutation).
- The continual-learning index at
.continual-learning/index.json(required, read before mutation). The index schema is a JSON object with an array of entries, each carryingfact,source_session,date, andcategory(one ofdecision,convention,constraint,resolved-problem,project-knowledge). - Update scope or focus filter (optional).
Procedure
- On a scheduled tick or watcher event, enumerate accessible prior chat transcripts and session logs in the local workspace. Done when: every accessible transcript and log is enumerated.
- Read the current AGENTS.md and
.continual-learning/index.jsonto establish the existing memory baseline. Done when: the existing memory baseline is read and the current set of recorded facts is known. - Extract candidate memory facts from the transcripts: decisions, conventions, constraints, resolved problems, and project-specific knowledge. Done when: candidate facts are extracted from every transcript.
- Deduplicate each candidate against the existing baseline; drop entries that duplicate, contradict without new evidence, or restate lower-signal information already recorded. Done when: every candidate is deduplicated against the baseline.
- Apply the high-signal gate. A candidate passes when it meets one of: records a decision that changed project direction, establishes a convention or constraint that governs future work, resolves a problem that recurred or is likely to recur, or captures project-specific knowledge not derivable from the codebase. Drop candidates that restate obvious or one-off information. Done when: every surviving candidate is classified and only high-signal entries remain.
- Capture the prior state of AGENTS.md and the index before writing, so the update can be rolled back. Apply the deduplicated high-signal updates to AGENTS.md and
.continual-learning/index.jsonas local writes only. Done when: the high-signal updates are written and the prior state is captured. - If no candidate survives deduplication and the gate, record an explicit no-update result. Done when: a no-update result is recorded or updates are applied.
What ships with it
1 file beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 2d ago Changed · -30 tokens per session bf137aaf9ed4
- 4d ago First seen · 46 lines · 66 tokens per session scan A f5fc5bb0fd1e
continual-learning is a skill published in the GitHub repository OutlineDriven/outline-driven-development (52 stars, last pushed 3d ago), licensed Apache-2.0. It adds 36 tokens to every session and 839 once invoked, about $0.0002 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-09-03.
Other skills, from other repositories
autolearn
Use when a verified non-trivial fix lands or existing solution docs need refresh. Not for unverified fixes.
consolidate-to-one-home
Fold a scattered fact into a single home and point the rest at it. Use when the same rule, constant, or definition is restated in several files, or the user says "consolidate this", "make this the single source of truth", or "ssotize this". The consolidation is proposed and approved before any edit lands. To…
memory-clean
Audit memory directory for structural issues (orphans, dangling refs, duplicates, missing sections, oversized entries) and staleness against session-history transcripts; report-first, fix-on-confirmation. Use when the user says "audit memory", "memory hygiene", or "find stale/duplicate memories".
memory-update
Scan agent's session-history transcripts for save-worthy signals (corrections, preferences, decisions, references), propose and write auto-memory files with valid frontmatter and MEMORY.md entry. Use when the user says "save this to memory", "remember that", or "scan this session for memories".
memory-sanitize
Produce share-safe copies of memory files under /tmp with PII redacted (paths, emails, session IDs, dates) and credentials scanned (tokens, keys); never mutates originals. Use when the user says "sanitize memory for sharing", "redact memory PII", or "scan memory for credentials".
shellm
Reference for the shellm system — recursive LLM shell, identity management, memory, skills, trajectory, and all CLI tools. Use when working on shellm itself, debugging agent behavior, or understanding how the pieces fit together.