Agent Workspace Practice MCP
An MCP server providing tools for agent operations architecture, including searching patterns, assessing workspace maturity, pricing context reads, and checking for data redaction.
Endpoint and access
No sign-in to connect and list tools. Individual tools may still need a vendor account, credentials, or permissions; check the documented prerequisites.
https://jamesross.ai/mcpWhat its tools do
6 tools, grouped by the hints the server advertises. Hints are self-declared, not verified behavior.
Can change data (1)
Checks and signals
Each line is separate evidence, not a trust score.
- Passed Endpoint answered the live check ·
- Warning Advertises resources/templates/list but does not implement it
- Availability 1 of 1 checks passed in 30 days
What you can do with it
- Search architecture patterns
- Assess workspace maturity
- Price context reads
- Check for redaction leaks
Source and maintenance
Protocol and capabilities
- Server
- jamesross-practice 1.3.1
- MCP protocol negotiated in check
- 2025-11-25
- Tools
- 6 advertised
- Resources
- 108 advertised
- Resource templates
- Not advertised
- Prompts
- Not advertised
{
"tools": {
"listChanged": false
},
"resources": {}
}Server-provided instructions
Untrusted server output shown for inspection. It is not Koskamo guidance and is excluded from knowledge search.
Tools
6 advertised
search_architecture — Search the published agent-workspace writing
Full-text search across the published architecture corpus: the 18 patterns and how they compose, the evaluation method, the seven learn-track modules, and the teardowns of other people's published agent systems. Returns ranked sections, each with its title, its source file, a snippet, and a link to the full text on GitHub. Use it when you want the passage itself rather than a pattern summary; lookup_pattern is the narrower tool that answers "which pattern covers this?". Every section is also readable in full as a resource under the architecture:// scheme (resources/list, resources/read).
{
"type": "object",
"required": [
"query"
],
"properties": {
"limit": {
"type": "integer",
"default": 5,
"maximum": 10,
"minimum": 1,
"description": "How many sections to return. 1-10, default 5."
},
"query": {
"type": "string",
"maxLength": 200,
"minLength": 1,
"description": "Search terms, e.g. \"canonical copy\", \"silent failure\", \"cache write cost\". Terms shorter than 3 characters are ignored."
}
},
"additionalProperties": false
}{
"readOnlyHint": true,
"openWorldHint": false
}lookup_pattern — Look up an agent-workspace pattern
Search the 18 architecture patterns behind this practice by keyword or by pattern number. Returns up to 3 matches, each with its number, title, the problem it opens on, and a link to the full text. Use it when you want the reasoning behind an agent-workspace design decision rather than a how-to.
{
"type": "object",
"required": [
"query"
],
"properties": {
"query": {
"type": "string",
"description": "Keywords to match against pattern titles and summaries, e.g. \"credentials\", \"silent failure\", \"context cost\". A bare number 1-18 looks that pattern up directly."
}
},
"additionalProperties": false
}{
"readOnlyHint": true,
"openWorldHint": false
}assess_workspace — Score an agent workspace or an organisation
Score an agent workspace across six dimensions: canonical knowledge, memory and context economics, verification and oversight, safety and permissions, telemetry and cost, provenance and delegation. Two variants: "practitioner" takes 18 answers (a single operator's workspace), "org" takes 12 (an organisation's agent readiness). Each answer is an integer 0-3 in question order, where 0 is the least mature option offered and 3 the most. Returns per-dimension scores, an overall percentage, a maturity band, and the two weakest dimensions with the gap that matters there. The question text lives at https://jamesross.ai/tools/maturity-check (practitioner) and https://jamesross.ai/tools/agent-readiness (org). Read it before answering on someone's behalf.
{
"type": "object",
"required": [
"variant",
"answers"
],
"properties": {
"answers": {
"type": "array",
"items": {
"type": "integer",
"maximum": 3,
"minimum": 0
},
"description": "Answers in question order: 18 integers for \"practitioner\", 12 for \"org\". Each is 0-3."
},
"variant": {
"enum": [
"practitioner",
"org"
],
"type": "string",
"description": "\"practitioner\" for the 18-question workspace check, \"org\" for the 12-question organisational readiness check."
}
},
"additionalProperties": false
}{
"readOnlyHint": true,
"openWorldHint": false
}price_context_read — Price a block of context by its position
Price a block of context by where it enters a session. The API is stateless, so the whole transcript is re-sent on every inference step, and a token costs its size times the number of turns left after it arrives. Returns the carried token volume, the cost of the block in USD, what the same read would have cost entering 20 turns from the end, and the multiplier between the two. Use it to decide whether a large read should move later in a session or into a subagent.
{
"type": "object",
"required": [
"size_tokens",
"entry_turn",
"total_turns"
],
"properties": {
"read_mult": {
"type": "number",
"default": 0.1,
"minimum": 0,
"description": "Cache-read price multiplier against input price. Default 0.1."
},
"entry_turn": {
"type": "number",
"minimum": 0,
"description": "Turn number at which the block enters the context (t)."
},
"write_mult": {
"type": "number",
"default": 1.25,
"minimum": 0,
"description": "Cache-write price multiplier against input price. Default 1.25."
},
"size_tokens": {
"type": "number",
"minimum": 0,
"description": "Size of the block in tokens (S)."
},
"total_turns": {
"type": "number",
"minimum": 0,
"description": "Total turns in the session (T). Must be at least entry_turn."
},
"price_in_per_mtok": {
"type": "number",
"default": 5,
"minimum": 0,
"description": "Input price in USD per million tokens. Default 5."
}
},
"additionalProperties": false
}{
"readOnlyHint": true,
"openWorldHint": false
}check_redaction — Pre-flight text for the shapes of private content
Scan text for the shapes of private content before it goes anywhere public: email addresses, absolute home paths, credential and token prefixes, private and link-local IPs, .local hostnames, and secret filenames. Returns a count per pattern class and a total, nothing else. The text you send and the substrings that matched are never stored, never logged and never echoed back. Only the counts are kept. A line carrying the marker "redaction-ok" is skipped. Accepts up to 100KB.
{
"type": "object",
"required": [
"text"
],
"properties": {
"text": {
"type": "string",
"description": "The text to scan. Up to 100KB; anything beyond that is truncated and the result says so."
}
},
"additionalProperties": false
}{
"readOnlyHint": true,
"openWorldHint": false
}request_capability — Request a capability this server does not have
Tell this server what capability you needed but didn't find here. Requests are logged and reviewed by a human; include no secrets or personal data.
{
"type": "object",
"required": [
"description"
],
"properties": {
"context": {
"type": "string",
"description": "Optional: what you were trying to do, and why the existing tools did not cover it. Up to 1KB."
},
"description": {
"type": "string",
"description": "What you needed. Up to 2KB."
}
},
"additionalProperties": false
}{
"readOnlyHint": false,
"openWorldHint": false,
"idempotentHint": false,
"destructiveHint": false
}Prompts
Not advertised
Resources
108 advertised
Patterns
PATTERNS.md
architecture://patterns--patterns1. Pure roles, composed with project facts
PATTERNS.md
architecture://patterns--1-pure-roles-composed-with-project-facts2. Classify-then-act, not ask-then-wait
PATTERNS.md
architecture://patterns--2-classify-then-act-not-ask-then-wait3. Make silent failure loud (the dead-man's switch)
PATTERNS.md
architecture://patterns--3-make-silent-failure-loud-the-dead-man-s-switch4. Tier by mechanical impact, not by tone
PATTERNS.md
architecture://patterns--4-tier-by-mechanical-impact-not-by-tone5. Memory points, it doesn't mirror
PATTERNS.md
architecture://patterns--5-memory-points-it-doesn-t-mirror6. Credentials live in one place, never in files
PATTERNS.md
architecture://patterns--6-credentials-live-in-one-place-never-in-files7. A cheap hook beats a careful agent
PATTERNS.md
architecture://patterns--7-a-cheap-hook-beats-a-careful-agent8. Audit the workspace like a fitness function
PATTERNS.md
architecture://patterns--8-audit-the-workspace-like-a-fitness-function9. Context is a budget, not a constant
PATTERNS.md
architecture://patterns--9-context-is-a-budget-not-a-constant10. A skill is editable weights — never adopt a self-edit without a gate
PATTERNS.md
architecture://patterns--10-a-skill-is-editable-weights-never-adopt-a-self-edit-without-a-gate11. A scaffold is a hypothesis — gate it behind a measurable signal
PATTERNS.md
architecture://patterns--11-a-scaffold-is-a-hypothesis-gate-it-behind-a-measurable-signal12. Loop selection: not everything should be a loop
PATTERNS.md
architecture://patterns--12-loop-selection-not-everything-should-be-a-loop13. Challenge half-formed ideas with a different lens — and hold a sample back to prove it helps
PATTERNS.md
architecture://patterns--13-challenge-half-formed-ideas-with-a-different-lens-and-hold-a-sample-back-to-prove-it-helps14. Delegation is a queue you fill, not work the agent finds
PATTERNS.md
architecture://patterns--14-delegation-is-a-queue-you-fill-not-work-the-agent-finds15. Price the lane before you migrate it
PATTERNS.md
architecture://patterns--15-price-the-lane-before-you-migrate-it16. A claim carries its provenance, or it is a guess
PATTERNS.md
architecture://patterns--16-a-claim-carries-its-provenance-or-it-is-a-guess17. One canonical copy, and pointers from everywhere else
PATTERNS.md
architecture://patterns--17-one-canonical-copy-and-pointers-from-everywhere-else18. Position is price — a token costs more the earlier you add it
PATTERNS.md
architecture://patterns--18-position-is-price-a-token-costs-more-the-earlier-you-add-itHow they compose
PATTERNS.md
architecture://patterns--how-they-composeEvaluation: knowing whether a workspace change helped
EVALUATION.md
architecture://evaluation--evaluation-knowing-whether-a-workspace-change-helpedThe instrument
EVALUATION.md
architecture://evaluation--the-instrument1. Deterministic checks only, never a judge
EVALUATION.md
architecture://evaluation--1-deterministic-checks-only-never-a-judge2. A pass-rate with variance, never a single number
EVALUATION.md
architecture://evaluation--2-a-pass-rate-with-variance-never-a-single-number3. Five fixture classes, and two of them are the point
EVALUATION.md
architecture://evaluation--3-five-fixture-classes-and-two-of-them-are-the-pointWhat it costs
EVALUATION.md
architecture://evaluation--what-it-costsHow it connects to the rest
EVALUATION.md
architecture://evaluation--how-it-connects-to-the-restWhere it lives
EVALUATION.md
architecture://evaluation--where-it-livesM0. Foundations — the session loop
learn/00-foundations.md
architecture://learn-00-foundations--m0-foundations-the-session-loopDo this
learn/00-foundations.md
architecture://learn-00-foundations--do-thisMeasure it
learn/00-foundations.md
architecture://learn-00-foundations--measure-itM1. Canonical knowledge — one source of truth
learn/01-canonical-knowledge.md
architecture://learn-01-canonical-knowledge--m1-canonical-knowledge-one-source-of-truthThe patterns
learn/01-canonical-knowledge.md
architecture://learn-01-canonical-knowledge--the-patternsDo this
learn/01-canonical-knowledge.md
architecture://learn-01-canonical-knowledge--do-thisMeasure it
learn/01-canonical-knowledge.md
architecture://learn-01-canonical-knowledge--measure-itM2. Context economics — context is spend
learn/02-context-economics.md
architecture://learn-02-context-economics--m2-context-economics-context-is-spendThe patterns
learn/02-context-economics.md
architecture://learn-02-context-economics--the-patternsDo this
learn/02-context-economics.md
architecture://learn-02-context-economics--do-thisMeasure it
learn/02-context-economics.md
architecture://learn-02-context-economics--measure-itM3. Verification & oversight — trust through checks
learn/03-verification-oversight.md
architecture://learn-03-verification-oversight--m3-verification-oversight-trust-through-checksThe patterns
learn/03-verification-oversight.md
architecture://learn-03-verification-oversight--the-patternsDo this
learn/03-verification-oversight.md
architecture://learn-03-verification-oversight--do-thisMeasure it
learn/03-verification-oversight.md
architecture://learn-03-verification-oversight--measure-itM4. Safety & permissions — cheap mechanical guards
learn/04-safety-permissions.md
architecture://learn-04-safety-permissions--m4-safety-permissions-cheap-mechanical-guardsThe patterns
learn/04-safety-permissions.md
architecture://learn-04-safety-permissions--the-patternsDo this
learn/04-safety-permissions.md
architecture://learn-04-safety-permissions--do-thisMeasure it
learn/04-safety-permissions.md
architecture://learn-04-safety-permissions--measure-itM5. Telemetry & cost — loud failure, priced lanes
learn/05-telemetry-cost.md
architecture://learn-05-telemetry-cost--m5-telemetry-cost-loud-failure-priced-lanesThe patterns
learn/05-telemetry-cost.md
architecture://learn-05-telemetry-cost--the-patternsDo this
learn/05-telemetry-cost.md
architecture://learn-05-telemetry-cost--do-thisMeasure it
learn/05-telemetry-cost.md
architecture://learn-05-telemetry-cost--measure-itM6. Provenance & delegation — sourced claims, mandated work
learn/06-provenance-delegation.md
architecture://learn-06-provenance-delegation--m6-provenance-delegation-sourced-claims-mandated-workThe patterns
learn/06-provenance-delegation.md
architecture://learn-06-provenance-delegation--the-patternsDo this
learn/06-provenance-delegation.md
architecture://learn-06-provenance-delegation--do-thisMeasure it
learn/06-provenance-delegation.md
architecture://learn-06-provenance-delegation--measure-itLearn the architecture
learn/README.md
architecture://learn-readme--learn-the-architectureWhere to start
learn/README.md
architecture://learn-readme--where-to-startThe modules
learn/README.md
architecture://learn-readme--the-modulesHow each module works
learn/README.md
architecture://learn-readme--how-each-module-worksTrack your progress
learn/README.md
architecture://learn-readme--track-your-progressQuestions and corrections
learn/README.md
architecture://learn-readme--questions-and-correctionsTeardown: 12-Factor Agents
teardowns/2026-08-27-12-factor-agents.md
architecture://teardowns-2026-08-27-12-factor-agents--teardown-12-factor-agentsWhat it is
teardowns/2026-08-27-12-factor-agents.md
architecture://teardowns-2026-08-27-12-factor-agents--what-it-isWhat works
teardowns/2026-08-27-12-factor-agents.md
architecture://teardowns-2026-08-27-12-factor-agents--what-worksThe trade-offs
teardowns/2026-08-27-12-factor-agents.md
architecture://teardowns-2026-08-27-12-factor-agents--the-trade-offsWhat's conspicuously absent
teardowns/2026-08-27-12-factor-agents.md
architecture://teardowns-2026-08-27-12-factor-agents--what-s-conspicuously-absentWhat this teaches
teardowns/2026-08-27-12-factor-agents.md
architecture://teardowns-2026-08-27-12-factor-agents--what-this-teachesCorrections
teardowns/2026-08-27-12-factor-agents.md
architecture://teardowns-2026-08-27-12-factor-agents--correctionsTeardown: herdr
teardowns/2026-08-28-herdr.md
architecture://teardowns-2026-08-28-herdr--teardown-herdrWhat it is
teardowns/2026-08-28-herdr.md
architecture://teardowns-2026-08-28-herdr--what-it-isWhat works
teardowns/2026-08-28-herdr.md
architecture://teardowns-2026-08-28-herdr--what-worksThe trade-offs
teardowns/2026-08-28-herdr.md
architecture://teardowns-2026-08-28-herdr--the-trade-offsWhat's conspicuously absent
teardowns/2026-08-28-herdr.md
architecture://teardowns-2026-08-28-herdr--what-s-conspicuously-absentWhat this teaches
teardowns/2026-08-28-herdr.md
architecture://teardowns-2026-08-28-herdr--what-this-teachesCorrections
teardowns/2026-08-28-herdr.md
architecture://teardowns-2026-08-28-herdr--correctionsTeardown: LifeOS
teardowns/2026-08-28-lifeos.md
architecture://teardowns-2026-08-28-lifeos--teardown-lifeosWhat it is
teardowns/2026-08-28-lifeos.md
architecture://teardowns-2026-08-28-lifeos--what-it-isWhat works
teardowns/2026-08-28-lifeos.md
architecture://teardowns-2026-08-28-lifeos--what-worksThe trade-offs
teardowns/2026-08-28-lifeos.md
architecture://teardowns-2026-08-28-lifeos--the-trade-offsWhat's conspicuously absent
teardowns/2026-08-28-lifeos.md
architecture://teardowns-2026-08-28-lifeos--what-s-conspicuously-absentWhat this teaches
teardowns/2026-08-28-lifeos.md
architecture://teardowns-2026-08-28-lifeos--what-this-teachesTeardown: DeepSeek Harness
teardowns/2026-09-05-deepseek-harness.md
architecture://teardowns-2026-09-05-deepseek-harness--teardown-deepseek-harnessWhat it is
teardowns/2026-09-05-deepseek-harness.md
architecture://teardowns-2026-09-05-deepseek-harness--what-it-isWhat works
teardowns/2026-09-05-deepseek-harness.md
architecture://teardowns-2026-09-05-deepseek-harness--what-worksThe trade-offs
teardowns/2026-09-05-deepseek-harness.md
architecture://teardowns-2026-09-05-deepseek-harness--the-trade-offsWhat's conspicuously absent
teardowns/2026-09-05-deepseek-harness.md
architecture://teardowns-2026-09-05-deepseek-harness--what-s-conspicuously-absentWhat this teaches
teardowns/2026-09-05-deepseek-harness.md
architecture://teardowns-2026-09-05-deepseek-harness--what-this-teachesTeardown: GPT-RAG
teardowns/2026-09-06-azure-gpt-rag.md
architecture://teardowns-2026-09-06-azure-gpt-rag--teardown-gpt-ragWhat it is
teardowns/2026-09-06-azure-gpt-rag.md
architecture://teardowns-2026-09-06-azure-gpt-rag--what-it-isWhat works
teardowns/2026-09-06-azure-gpt-rag.md
architecture://teardowns-2026-09-06-azure-gpt-rag--what-worksThe trade-offs
teardowns/2026-09-06-azure-gpt-rag.md
architecture://teardowns-2026-09-06-azure-gpt-rag--the-trade-offsWhat's conspicuously absent
teardowns/2026-09-06-azure-gpt-rag.md
architecture://teardowns-2026-09-06-azure-gpt-rag--what-s-conspicuously-absentWhat this teaches
teardowns/2026-09-06-azure-gpt-rag.md
architecture://teardowns-2026-09-06-azure-gpt-rag--what-this-teachesCorrections
teardowns/2026-09-06-azure-gpt-rag.md
architecture://teardowns-2026-09-06-azure-gpt-rag--correctionsTeardown: OpenHarness
teardowns/2026-10-02-openharness.md
architecture://teardowns-2026-10-02-openharness--teardown-openharnessWhat it is
teardowns/2026-10-02-openharness.md
architecture://teardowns-2026-10-02-openharness--what-it-isWhat works
teardowns/2026-10-02-openharness.md
architecture://teardowns-2026-10-02-openharness--what-worksThe trade-offs
teardowns/2026-10-02-openharness.md
architecture://teardowns-2026-10-02-openharness--the-trade-offsWhat's conspicuously absent
teardowns/2026-10-02-openharness.md
architecture://teardowns-2026-10-02-openharness--what-s-conspicuously-absentWhat this teaches
teardowns/2026-10-02-openharness.md
architecture://teardowns-2026-10-02-openharness--what-this-teachesTeardowns
teardowns/README.md
architecture://teardowns-readme--teardownsWhy teardowns
teardowns/README.md
architecture://teardowns-readme--why-teardownsGround rules
teardowns/README.md
architecture://teardowns-readme--ground-rulesPage conventions
teardowns/README.md
architecture://teardowns-readme--page-conventionsDistribution
teardowns/README.md
architecture://teardowns-readme--distributionWhat six readings show
teardowns/README.md
architecture://teardowns-readme--what-six-readings-showPublished
teardowns/README.md
architecture://teardowns-readme--publishedDisagree with a reading?
teardowns/README.md
architecture://teardowns-readme--disagree-with-a-reading30-day availability
Aggregated from read-only remote checks; this is not a tool-execution or security guarantee.
- Successful checks
- 1 of 1 (100%)
- Average latency
- 510 ms
- 95th percentile latency
- 510 ms
Maintenance history
Public check and source-refresh events. Editorial and ownership records remain private.
- · probe · passed · 510 ms · metadata changed · 6 tool(s), 108 resource(s); warnings: Advertises resources/templates/list but does not implement it
Connect from your client
Each snippet registers the same endpoint; pick the client you use.
Claude Code
claude mcp add --transport http agent-workspace-practice-mcp https://jamesross.ai/mcpCodex
codex mcp add agent-workspace-practice-mcp --url https://jamesross.ai/mcpCursor
{
"mcpServers": {
"agent-workspace-practice-mcp": {
"url": "https://jamesross.ai/mcp"
}
}
}Cline
{
"mcpServers": {
"agent-workspace-practice-mcp": {
"type": "streamableHttp",
"url": "https://jamesross.ai/mcp",
"disabled": false
}
}
}VS Code
{
"servers": {
"agent-workspace-practice-mcp": {
"type": "http",
"url": "https://jamesross.ai/mcp"
}
}
}Gemini CLI
gemini mcp add --transport http agent-workspace-practice-mcp https://jamesross.ai/mcpClaude Desktop
https://jamesross.ai/mcpWindsurf
{
"mcpServers": {
"agent-workspace-practice-mcp": {
"serverUrl": "https://jamesross.ai/mcp"
}
}
}Let your agent find its tools
Find servers like Agent Workspace Practice MCP from your agent. Add Koskamo once to search the index for endpoints, access requirements, and client setup. Public discovery needs no key.