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 agentmods add skills/msinclair25/mdevolved/mdevolved-albatrossnpx skills add msinclair25/mdevolved --skill mdevolved-albatrossgit clone --depth 1 https://github.com/msinclair25/mdevolvedWhat 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 | $0.00036 | $0.00507 |
| Opus 5 | $0.00018 | $0.00253 |
| Sonnet 5 | $0.00007 | $0.00101 |
| Haiku 4.5 | $0.00004 | $0.00051 |
Grade A, and why
mdevolved-albatross 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 yesterday.
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.
What it actually says
MDevolved + Albatross
Use this pack only for Albatross-side setup. MDevolved remains the universal
remote MCP server and the authority boundary; Albatross still owns its local
agents, tools, workspaces, retries, and /reset behavior.
Setup
- Preserve every existing Albatross setting and MCP server.
- Use the exact HTTPS
/mcpURL (HTTP only for localhost development). - Add the generated
mdevolvedentry toagent.config.jsonusing the pinnedmcp-remotebridge, then run/mcp trust mdevolved. - Merge the marked profile into
.albatross/prompt.md; preserve everything outside the marked block. - Use a separate participant ID and OAuth authorization only for an independently attributable writer or reviewer. A label never grants access.
Connect or resume
Read .mdevolvedignore first. If it is absent, read .owdignore as a legacy
fallback. If both exist and disagree, stop and ask the owner to resolve the
conflict; never guess or create a second Project.
With a receipt, call mcp__mdevolved__mdevolved_resume first with its exact
Project UUID and policy. Without one, call mcp__mdevolved__connection_info,
then mcp__mdevolved__open_project. If approval is pending, repeat only the
same mcp__mdevolved__wait_for_project_connection request with its exact key.
Until resume returns localVaultAccess, the writer role is unconfirmed.
After /reset, resume the same receipt before local writes. Submit only cited,
useful results to MDevolved; Albatross transcripts, checkpoints, and private
runtime state remain local.
Before any shell, skill, or filesystem write, obey the returned
localVaultAccess.role. Project submissions never grant local write authority.
Legacy compatibility
Existing pre-MD9 owd server/tool configuration and .owdignore receipts
remain readable. Preserve them and do not silently re-authorize or delete them;
new setup should use the canonical mdevolved names and .mdevolvedignore.
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.
- yesterday First seen · 47 lines · 36 tokens per session scan A e518ebb5b5e2
mdevolved-albatross is a skill published in the GitHub repository msinclair25/mdevolved (2 stars, last pushed yesterday), licensed Apache-2.0. It adds 36 tokens to every session and 507 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 2026-09-01.
Other skills, from other repositories
backups
Use when designing or auditing a backup-and-restore program that must survive a real disaster: setting defensible RPO/RTO targets, laying out 3-2-1-1-0 copies that are offsite and immutable, wiring point-in-time recovery, and proving restores work on a schedule. NOT tuning Postgres internals or writing the…
tfw-identity
Внутренний fail-closed механизм профилей и локальной привязки Assisted; отделяет участника, роли, task owner и роль ИИ.
tfw-init
Command /tfw-init initializes Trace-First Workflow or attaches and repairs TFW tooling in an existing project. Use for /tfw-init, first-time setup, project discovery, adapter installation, or Codex adapter repair.
tfw-update
Command /tfw-update upgrades and synchronizes local Trace-First Workflow files and adapters while preserving project state. Use for /tfw-update, upstream upgrades, version comparison, migration, or adapter repair.
tfw-config
Command /tfw-config audits or changes Trace-First Workflow configuration and propagates values to every registered inline location. Use for /tfw-config, config verification, projectconfig.yaml changes, or Config Sync Registry updates.
tfw-docs
Command /tfw-docs updates Trace-First Workflow project documentation and technical debt after review. Use for /tfw-docs or updates to KNOWLEDGE.md sections 1-3 or TECHDEBT.md from RF/REVIEW results.