create-academic-research 0.1.5 → 0.1.7

package/README.md

@@ -46,11 +46,16 @@ By default, the wizard:
 
 - creates the repository structure;
 - installs the project-local `VincenzoImp/academic-research-skills` package;
-- enables the scholarly MCP records for `arxiv`, `semantic-scholar`, and
-  `openalex`;
+- enables only the low-friction `arxiv` MCP record;
+- documents the wider MCP catalog, including Semantic Scholar, OpenAlex, DBLP,
+  PubMed, Zotero, Crossref, Overleaf, and fallback aggregators;
 - writes `configs/capabilities.yaml`;
 - writes `docs/agent/capability-profile.md`;
+- writes `docs/agent/mcp-setup.md`;
 - writes `docs/agent/generated/mcp.json` unless an explicit agent target is set;
+- includes `scripts/README.md` for repeatable command entrypoints;
+- includes reusable `wiki/templates/` pages for sources, claims, experiments,
+  decisions, reviewer concerns, and research questions;
 - appends the onboarding event to `wiki/log.md`;
 - does not install external MCP tools unless explicitly requested.
 
@@ -58,6 +63,12 @@ Use `--preset enhanced` when you also want the curated complementary external
 skill bundles for agent engineering, frontend work, testing, document formats,
 and PDF conversion.
 
+Research-specific skills are intentionally not pulled from external packages by
+default. The academic research workflow, literature review, citation audit,
+paper writing, peer review, rebuttal, and reproduction policies come from
+`VincenzoImp/academic-research-skills`. External skills installed by the wizard
+are complementary tooling only.
+
 ## Non-Interactive Create
 
 ```bash
@@ -76,6 +87,7 @@ Inside a generated project:
 
 ```bash
 npx academic-research doctor
+npx academic-research setup
 npx academic-research rename --title "New Title" --slug new-title --package new_title
 npx academic-research agents list
 npx academic-research skills presets
@@ -90,15 +102,21 @@ npx academic-research mcp list
 npx academic-research mcp enabled
 npx academic-research mcp available
 npx academic-research mcp commands arxiv
-npx academic-research mcp enable arxiv openalex
+npx academic-research mcp env openalex semantic-scholar zotero
+npx academic-research mcp enable arxiv dblp
 npx academic-research mcp disable arxiv
 npx academic-research mcp install arxiv
 npx academic-research mcp uninstall arxiv
+npx academic-research mcp smoke
 npx academic-research mcp doctor
 ```
 
 ## Command Model
 
+`academic-research setup` is a non-destructive onboarding status command. It
+prints the active preset, agent, skill counts, enabled MCP records, and next
+commands without changing files.
+
 Skills are project-local by default.
 
 | Command | Meaning |
@@ -118,11 +136,13 @@ MCP commands are split by side-effect:
 | `mcp enabled` | List only enabled MCP server ids. |
 | `mcp available` | List the local MCP catalog. |
 | `mcp commands` | Print finite external install commands without running them. Runtime-only `uvx`/`npx` servers may have no install command. |
+| `mcp env` | Print required/recommended env vars, hosted endpoints, local prerequisites, and setup commands for selected servers. |
 | `mcp enable` | Enable an MCP server in project records and generated snippets. |
 | `mcp disable` | Remove an MCP server from project records and generated snippets. |
 | `mcp install` | Run finite external tool install commands for selected MCP servers. It must not launch stdio MCP servers. |
 | `mcp uninstall` | Run the external uninstall command when one exists. |
-| `mcp doctor` | Validate enabled MCP records and generated snippets. |
+| `mcp smoke` | Print non-launching readiness diagnostics for enabled or selected MCP servers. |
+| `mcp doctor` | Validate enabled MCP records, generated snippets, required env vars, and documented manual prerequisites. |
 
 ## Companion Skills
 
@@ -147,11 +167,37 @@ Preset intent:
 | Preset | Intent |
 |---|---|
 | `minimal` | Academic research skills only, no MCP records. |
-| `default` | Academic research skills plus core scholarly MCP records. |
+| `default` | Academic research skills plus the low-friction arXiv MCP record. |
 | `enhanced` | `default` plus curated external complementary skill bundles. |
-| `literature` | SOTA and systematic-review work with citation-library MCP records. |
-| `writing` | Paper-writing and Overleaf-oriented work. |
-| `full` | Broad optional connector and specialist setup. |
+| `literature` | SOTA and systematic-review work with arXiv plus DBLP for computer science bibliography. |
+| `writing` | Paper-writing work; Overleaf is documented as an opt-in credentialed integration. |
+| `full` | Broad setup with low-friction arXiv and DBLP records plus the full optional MCP catalog documented. |
+
+MCP defaults are intentionally conservative. Semantic Scholar, OpenAlex,
+Zotero, Overleaf, Crossref, and fallback aggregators are useful, but they need
+API keys, local apps, manual setup, or source-policy review. Enable them with
+`npx academic-research mcp enable <server>` after reading
+`docs/agent/mcp-setup.md`, use `npx academic-research mcp env <server>` to see
+runtime prerequisites, then run `npx academic-research mcp doctor`.
+
+The MCP catalog distinguishes local runtime adapters from hosted endpoints and
+manual integrations. arXiv and DBLP are low-friction local `uvx` runtimes.
+Semantic Scholar is useful for citation graphs but works best with
+`SEMANTIC_SCHOLAR_API_KEY`. OpenAlex requires `OPENALEX_API_KEY` for the
+selected local adapter; OpenAlex keys are free for normal academic use with
+daily free usage. PubMed is a biomedical-specific `npx` runtime and remains
+opt-in. Zotero needs the local Zotero desktop app and Zoty setup. Overleaf is
+manual and credentialed. Crossref and broad paper-search aggregators are kept
+as fallback/manual entries until a project explicitly needs them.
+
+Generated MCP snippets are project documentation and client-ready config, not
+live tools by themselves. Your MCP client must load the generated snippet, and
+the referenced commands must be available on your machine or runnable through
+`uvx`/`npx`. `mcp install` only runs finite setup commands such as the arXiv
+tool install; it deliberately does not launch stdio MCP servers.
+Use `mcp smoke` for a non-launching readiness pass before wiring a client: it
+checks required env vars, manual/local-service status, and whether runtime
+commands are visible on `PATH`.
 
 ## Validate This Package
 
@@ -169,8 +215,8 @@ Releases are tag-driven. Update `package.json` and `package-lock.json`, commit
 the change, create `vX.Y.Z`, and push the tag:
 
 ```bash
-git tag -a v0.1.5 -m "v0.1.5"
-git push origin main v0.1.5
+git tag -a v0.1.7 -m "v0.1.7"
+git push origin main v0.1.7
 ```
 
 Once the GitHub repository is public, the release workflow validates the tag

package/dist/src/capabilities.js

@@ -25,6 +25,7 @@ export async function writeCapabilities(root, state) {
     };
     await writeFile(join(root, "configs/capabilities.yaml"), YAML.stringify(next), "utf8");
     await writeCapabilityProfile(root, next);
+    await writeMcpSetup(root, next);
     await writeMcpSnippet(root, next);
     await appendCapabilityLog(root, next);
 }
@@ -233,6 +234,19 @@ export async function doctorMcpServers(root) {
         const server = AGENT_STACK.mcp_servers[name];
         if (!server)
             continue;
+        for (const envName of server.required_env) {
+            if (!process.env[envName]) {
+                errors.push(`${name}: missing required environment variable: ${envName}`);
+            }
+        }
+        for (const envName of server.recommended_env) {
+            if (!process.env[envName]) {
+                warnings.push(`${name}: recommended environment variable not set: ${envName}`);
+            }
+        }
+        if (server.local_service) {
+            warnings.push(`${name}: requires local service: ${server.local_service}`);
+        }
         if (!server.command) {
             warnings.push(`${name}: manual setup only; no generated client command`);
             continue;
@@ -240,9 +254,6 @@ export async function doctorMcpServers(root) {
         if (!generatedServers.has(name)) {
             errors.push(`${name}: enabled but missing from generated MCP snippet`);
         }
-        if (!server.install_command) {
-            warnings.push(`${name}: no automated install command is defined`);
-        }
     }
     return { ok: errors.length === 0, errors, warnings, enabled };
 }
@@ -290,13 +301,46 @@ async function writeCapabilityProfile(root, state) {
     else {
         for (const name of state.mcp_servers) {
             const server = AGENT_STACK.mcp_servers[name];
-            const status = server?.command ? "generated config" : "manual setup";
+            const status = server?.command ? server.execution_mode : "manual setup";
             lines.push(`- \`${name}\` (${status}): ${server?.smoke_test ?? "Smoke-test before use."}`);
         }
     }
     lines.push("", "## Rules", "", "- Skill installation is project-local by default.", "- Agent target `universal` installs one shared project-local `.agents/skills` copy.", "- MCP enable/disable changes project records; install/uninstall changes external tools.", "- Keep API keys, tokens, cookies, and browser sessions out of git.", "- Cite repository source records, not raw MCP output alone.", "");
     await writeFile(join(root, "docs/agent/capability-profile.md"), lines.join("\n"), "utf8");
 }
+async function writeMcpSetup(root, state) {
+    const enabled = new Set(state.mcp_servers ?? []);
+    const lines = [
+        "# MCP Setup",
+        "",
+        "This file is generated from the project-local academic research capability stack.",
+        "MCP records are configuration snippets and setup guidance; the active MCP client must load the generated snippet before these servers become live tools.",
+        "",
+        "## Enabled MCP Servers",
+        ""
+    ];
+    if (enabled.size === 0) {
+        lines.push("- None.");
+    }
+    else {
+        for (const name of state.mcp_servers) {
+            const server = AGENT_STACK.mcp_servers[name];
+            if (!server)
+                continue;
+            lines.push(`- \`${name}\` (${server.readiness}, ${server.priority}): ${server.source_need}`, `  - Source: \`${server.source}\``, `  - Execution mode: \`${server.execution_mode}\``, ...(server.hosted_url ? [`  - Hosted endpoint: <${server.hosted_url}>`] : []), `  - Runtime: ${formatRuntime(server.command, server.args)}`, `  - Install command: ${server.install_command ? `\`${server.install_command}\`` : "none; runtime-only or manual setup"}`, ...server.setup_commands.map((command) => `  - Setup command: \`${command}\``), `  - Smoke test: ${server.smoke_test}`, `  - Risks: ${server.risks}`);
+            appendMcpPrerequisiteLines(lines, server.required_env, server.recommended_env, server.local_service);
+        }
+    }
+    lines.push("", "## Available MCP Catalog", "");
+    for (const [name, server] of Object.entries(AGENT_STACK.mcp_servers)) {
+        const status = enabled.has(name) ? "enabled" : "available";
+        lines.push(`- \`${name}\` (${status}, ${server.readiness}, ${server.priority}): ${server.source_need}`, `  - Source: \`${server.source}\``, `  - Execution mode: \`${server.execution_mode}\``, ...(server.hosted_url ? [`  - Hosted endpoint: <${server.hosted_url}>`] : []), ...server.setup_commands.map((command) => `  - Setup command: \`${command}\``));
+        appendMcpPrerequisiteLines(lines, server.required_env, server.recommended_env, server.local_service);
+    }
+    lines.push("", "## Operating Rules", "", "- Keep secrets in your shell, MCP client secret store, or local untracked files; do not commit tokens or API keys.", "- Prefer the smallest enabled MCP set that covers the current research question.", "- Treat MCP output as retrieval metadata. Promote claims into repository source records only after source ingestion and citation audit.", "- Run `npx academic-research mcp doctor` after changing MCP records or environment variables.", "");
+    await mkdir(join(root, "docs/agent"), { recursive: true });
+    await writeFile(join(root, "docs/agent/mcp-setup.md"), lines.join("\n"), "utf8");
+}
 async function appendCapabilityLog(root, state) {
     const logPath = join(root, "wiki/log.md");
     const date = new Date().toISOString().slice(0, 10);
@@ -311,6 +355,20 @@ function renderSkillCommand(command, agent) {
     const agentFlag = normalized === AUTO_AGENT ? "" : `--agent '${normalized}'`;
     return command.replaceAll("{agent_flag}", agentFlag).replaceAll("{agent}", normalized);
 }
+function appendMcpPrerequisiteLines(lines, requiredEnv, recommendedEnv, localService) {
+    if (requiredEnv.length > 0)
+        lines.push(`  - Requires env: ${requiredEnv.map((name) => `\`${name}\``).join(", ")}`);
+    if (recommendedEnv.length > 0) {
+        lines.push(`  - Recommended env: ${recommendedEnv.map((name) => `\`${name}\``).join(", ")}`);
+    }
+    if (localService)
+        lines.push(`  - Local prerequisite: ${localService}`);
+}
+function formatRuntime(command, args) {
+    if (!command)
+        return "manual setup";
+    return `\`${[command, ...args].join(" ")}\``;
+}
 async function readCapabilitiesFile(root) {
     const path = join(root, "configs/capabilities.yaml");
     return normalizeCapabilityState(YAML.parse(await readFile(path, "utf8")));

package/dist/src/cli.js

@@ -1,7 +1,7 @@
-import { readFileSync } from "node:fs";
-import { basename, dirname, join, resolve } from "node:path";
+import { existsSync, readFileSync } from "node:fs";
+import { basename, delimiter, dirname, join, resolve } from "node:path";
 import { fileURLToPath } from "node:url";
-import { disableMcpServers, doctorMcpServers, enableMcpServers, DEFAULT_AGENT, installMcpTools, installSkills, listInstalledSkills, mcpToolCommandTexts, readCapabilities, removeSkills, uninstallMcpTools, updateSkills } from "./capabilities.js";
+import { assertKnownMcpServers, disableMcpServers, doctorMcpServers, enableMcpServers, DEFAULT_AGENT, installMcpTools, installSkills, listInstalledSkills, mcpToolCommandTexts, readCapabilities, removeSkills, uninstallMcpTools, updateSkills } from "./capabilities.js";
 import { createProject, doctorProject, renameProject } from "./project.js";
 import { askCreateOptions } from "./prompts.js";
 import { AGENT_STACK, presetMcpServers } from "./stack.js";
@@ -100,6 +100,8 @@ async function lifecycleMain(argv) {
     }
     if (command === "doctor")
         return doctorCommand(argv.slice(1));
+    if (command === "setup")
+        return setupCommand(argv.slice(1));
     if (command === "rename")
         return renameCommand(argv.slice(1));
     if (command === "agents")
@@ -125,6 +127,41 @@ async function doctorCommand(argv) {
         console.log(`OK: ${root}`);
     return result.ok ? 0 : 1;
 }
+async function setupCommand(argv) {
+    const parsed = parseFlags(argv, ROOT_FLAGS);
+    if (flagBool(parsed.flags, "help")) {
+        printSetupHelp();
+        return 0;
+    }
+    assertNoArguments(parsed.positionals, "setup");
+    const root = resolve(flagString(parsed.flags, "root") ?? ".");
+    const project = await doctorProject(root);
+    const state = await readCapabilities(root);
+    const skills = await listInstalledSkills(root);
+    const skillIds = new Set(skills.map((skill) => skill.name));
+    console.log("Project Setup");
+    console.log(`root\t${root}`);
+    console.log(`doctor\t${project.ok ? "ok" : "error"}`);
+    console.log(`agent\t${state.agent}`);
+    console.log(`preset\t${state.preset}`);
+    console.log(`scope\t${state.scope}`);
+    console.log(`installed_skill_ids\t${skillIds.size}`);
+    console.log(`installed_skill_copies\t${skills.length}`);
+    console.log(`mcp_enabled\t${state.mcp_servers.length > 0 ? state.mcp_servers.join(",") : "none"}`);
+    if (!project.ok) {
+        for (const error of project.errors)
+            console.error(`ERROR: ${error}`);
+    }
+    console.log("");
+    console.log("Next Commands");
+    console.log(`academic-research skills install --preset ${state.preset}`);
+    console.log("academic-research skills status");
+    console.log("academic-research mcp list");
+    console.log("academic-research mcp env");
+    console.log("academic-research mcp smoke");
+    console.log("academic-research doctor");
+    return project.ok ? 0 : 1;
+}
 async function renameCommand(argv) {
     const parsed = parseFlags(argv, RENAME_FLAGS);
     if (flagBool(parsed.flags, "help")) {
@@ -244,10 +281,10 @@ async function mcpCommand(argv) {
         assertNoArguments(parsed.positionals, "mcp list");
         const state = await readCapabilities(root);
         const enabled = new Set(state.mcp_servers ?? []);
+        console.log("status\tid\treadiness\texecution_mode\tdescription");
         for (const [name, server] of Object.entries(AGENT_STACK.mcp_servers)) {
             const status = enabled.has(name) ? "enabled" : "available";
-            const installer = mcpInstallMode(server.install_command, server.command);
-            console.log(`${status}\t${name}\t${server.source_need}\t${installer}`);
+            console.log(`${status}\t${name}\t${server.readiness}\t${server.execution_mode}\t${server.source_need}`);
         }
         return 0;
     }
@@ -263,9 +300,9 @@ async function mcpCommand(argv) {
     if (subcommand === "available") {
         assertOnlyOptions(parsed.flags, "mcp available", []);
         assertNoArguments(parsed.positionals, "mcp available");
+        console.log("id\treadiness\texecution_mode\tdescription");
         for (const [name, server] of Object.entries(AGENT_STACK.mcp_servers)) {
-            const installer = mcpInstallMode(server.install_command, server.command);
-            console.log(`${name}\t${server.source_need}\t${installer}`);
+            console.log(`${name}\t${server.readiness}\t${server.execution_mode}\t${server.source_need}`);
         }
         return 0;
     }
@@ -278,6 +315,15 @@ async function mcpCommand(argv) {
             console.log(command);
         return 0;
     }
+    if (subcommand === "env") {
+        assertOnlyOptions(parsed.flags, "mcp env", ["root"]);
+        const root = resolve(flagString(parsed.flags, "root") ?? ".");
+        const selected = parsed.positionals.length > 0 ? parsed.positionals : (await readCapabilities(root)).mcp_servers;
+        assertKnownMcpServers(selected);
+        console.log("id\ttype\tvalue");
+        printMcpEnvironment(selected);
+        return 0;
+    }
     if (subcommand === "enable") {
         assertOnlyOptions(parsed.flags, "mcp enable", ["root", "agent"]);
         const root = resolve(flagString(parsed.flags, "root") ?? ".");
@@ -308,6 +354,24 @@ async function mcpCommand(argv) {
         console.log(`Ran ${result.count ?? 0} MCP uninstall command(s).`);
         return 0;
     }
+    if (subcommand === "smoke") {
+        assertOnlyOptions(parsed.flags, "mcp smoke", ["root"]);
+        const root = resolve(flagString(parsed.flags, "root") ?? ".");
+        const state = await readCapabilities(root);
+        const explicitSelection = parsed.positionals.length > 0;
+        const selected = explicitSelection ? parsed.positionals : state.mcp_servers;
+        assertKnownMcpServers(selected);
+        const failed = printMcpSmokeDiagnostics(selected);
+        if (!explicitSelection) {
+            const result = await doctorMcpServers(root);
+            for (const error of result.errors)
+                console.error(`ERROR: ${error}`);
+            for (const warning of result.warnings)
+                console.warn(`WARN: ${warning}`);
+            return result.ok && !failed ? 0 : 1;
+        }
+        return failed ? 1 : 0;
+    }
     if (subcommand === "doctor") {
         assertOnlyOptions(parsed.flags, "mcp doctor", ["root"]);
         const root = resolve(flagString(parsed.flags, "root") ?? ".");
@@ -405,11 +469,6 @@ function assertOnlyOptions(flags, command, allowedOptions) {
         throw new Error(`${command} does not accept ${unexpected.map((name) => `--${name}`).join(", ")}`);
     }
 }
-function mcpInstallMode(installCommand, runtimeCommand) {
-    if (installCommand)
-        return installCommand;
-    return runtimeCommand ? "runtime-only" : "manual";
-}
 export function formatInteractiveCreateGuide() {
     const presetLines = Object.entries(AGENT_STACK.presets).map(([name, preset]) => `  ${name.padEnd(10)} ${preset.description}`);
     return [
@@ -432,8 +491,11 @@ export function formatInteractiveCreateGuide() {
         "Skill and MCP behavior:",
         "  Skills are copied into the project, not installed globally.",
         "  MCP records are written into configs/capabilities.yaml and docs/agent/generated/.",
+        "  docs/agent/mcp-setup.md records enabled servers, optional catalog entries, env vars, and smoke tests.",
+        "  default enables only low-friction arXiv; credentialed/local services are opt-in.",
         "  MCP installers are optional and run only finite installer commands.",
-        "  runtime-only MCP servers are configured for the MCP client but have no install step.",
+        "  MCP execution modes are explicit: uvx-runtime, npx-runtime, local-service, manual, or fallback.",
+        "  Use `academic-research mcp env <server>` to inspect env vars and local prerequisites.",
         ""
     ].join("\n");
 }
@@ -469,7 +531,7 @@ function printMissingTargetHelp() {
 }
 function printLifecycleHelp() {
     console.log([
-        "Usage: academic-research <doctor|rename|agents|skills|mcp>",
+        "Usage: academic-research <doctor|setup|rename|agents|skills|mcp>",
         "",
         "Manage a generated academic research repository after creation.",
         "",
@@ -478,6 +540,17 @@ function printLifecycleHelp() {
         "  -v, --version            Show package version."
     ].join("\n"));
 }
+function printSetupHelp() {
+    console.log([
+        "Usage: academic-research setup [options]",
+        "",
+        "Print project onboarding status and next commands without changing files.",
+        "",
+        "Options:",
+        "  --root <path>            Project root. Default: current directory.",
+        "  -h, --help               Show this help."
+    ].join("\n"));
+}
 function printAgentsHelp() {
     console.log([
         "Usage: academic-research agents <list>",
@@ -508,9 +581,9 @@ function printSkillsHelp() {
 }
 function printMcpHelp() {
     console.log([
-        "Usage: academic-research mcp <list|enabled|available|commands|enable|disable|install|uninstall|doctor> [servers...]",
+        "Usage: academic-research mcp <list|enabled|available|commands|env|enable|disable|install|uninstall|smoke|doctor> [servers...]",
         "",
-        "Manage MCP records and finite external MCP tool installs.",
+        "Manage MCP records, readiness checks, and finite external MCP tool installs.",
         "",
         "Options:",
         "  --root <path>            Project root for project-state commands.",
@@ -518,6 +591,79 @@ function printMcpHelp() {
         "  -h, --help               Show this help."
     ].join("\n"));
 }
+function printMcpSmokeDiagnostics(servers) {
+    let failed = false;
+    console.log("id\tstatus\truntime\tcheck");
+    for (const name of servers) {
+        const server = AGENT_STACK.mcp_servers[name];
+        const missingRequired = server.required_env.filter((envName) => !process.env[envName]);
+        if (missingRequired.length > 0)
+            failed = true;
+        const runtime = server.command ? [server.command, ...server.args].join(" ") : "manual setup";
+        let status = "manual";
+        if (missingRequired.length > 0) {
+            status = `missing-required-env:${missingRequired.join(",")}`;
+        }
+        else if (server.command && commandExists(server.command)) {
+            status = "runtime-found";
+        }
+        else if (server.command) {
+            status = "runtime-missing";
+        }
+        else if (server.local_service) {
+            status = "manual-local-service";
+        }
+        console.log(`${name}\t${status}\t${runtime}\t${server.smoke_test}`);
+    }
+    return failed;
+}
+function printMcpEnvironment(servers) {
+    for (const name of servers) {
+        const server = AGENT_STACK.mcp_servers[name];
+        let wroteLine = false;
+        for (const envName of server.required_env) {
+            console.log(`${name}\trequired\t${envName}`);
+            wroteLine = true;
+        }
+        for (const envName of server.recommended_env) {
+            console.log(`${name}\trecommended\t${envName}`);
+            wroteLine = true;
+        }
+        if (server.hosted_url) {
+            console.log(`${name}\thosted-endpoint\t${server.hosted_url}`);
+            wroteLine = true;
+        }
+        if (server.local_service) {
+            console.log(`${name}\tlocal-service\t${server.local_service}`);
+            wroteLine = true;
+        }
+        for (const command of server.setup_commands) {
+            console.log(`${name}\tsetup-command\t${command}`);
+            wroteLine = true;
+        }
+        if (!wroteLine)
+            console.log(`${name}\tnone\t-`);
+    }
+}
+function commandExists(command) {
+    if (!command)
+        return false;
+    if (command.includes("/") || command.includes("\\"))
+        return existsSync(command);
+    const pathValue = process.env.PATH ?? "";
+    const extensions = process.platform === "win32"
+        ? (process.env.PATHEXT ?? ".EXE;.CMD;.BAT;.COM").split(";")
+        : [""];
+    for (const directory of pathValue.split(delimiter).filter(Boolean)) {
+        for (const extension of extensions) {
+            const hasExtension = extension && command.toLowerCase().endsWith(extension.toLowerCase());
+            const candidate = join(directory, hasExtension ? command : `${command}${extension}`);
+            if (existsSync(candidate))
+                return true;
+        }
+    }
+    return false;
+}
 function readPackageVersion() {
     try {
         const packageJson = JSON.parse(readFileSync(join(packageRoot, "package.json"), "utf8"));

package/dist/src/project.js

@@ -136,11 +136,19 @@ export async function doctorProject(root) {
         "configs/agent-stack.yaml",
         "configs/capabilities.yaml",
         "docs/agent/capability-profile.md",
+        "docs/agent/mcp-setup.md",
         "docs/agent/generated",
+        "scripts/README.md",
         "sources/source-ledger.csv",
         "sota/literature-matrix.csv",
         "wiki/index.md",
-        "wiki/log.md"
+        "wiki/log.md",
+        "wiki/templates/source-page.md",
+        "wiki/templates/claim-page.md",
+        "wiki/templates/experiment-page.md",
+        "wiki/templates/decision-record.md",
+        "wiki/templates/reviewer-concern.md",
+        "wiki/templates/research-question.md"
     ];
     for (const relative of required) {
         if (!(await exists(join(target, relative))))
@@ -206,7 +214,7 @@ async function writeGeneratedPackageJson(root, { slug }) {
     const path = join(root, "package.json");
     const data = await readJson(path);
     const existingSpec = data.devDependencies?.["create-academic-research"];
-    const packageSpec = process.env.CREATE_ACADEMIC_RESEARCH_PACKAGE_SPEC ?? existingSpec ?? "0.1.5";
+    const packageSpec = process.env.CREATE_ACADEMIC_RESEARCH_PACKAGE_SPEC ?? existingSpec ?? "0.1.7";
     data.name = slug;
     data.devDependencies = {
         ...(data.devDependencies ?? {}),

package/dist/src/prompts.js

@@ -9,7 +9,8 @@ export async function askCreateOptions(defaults, locks = {}) {
         const preset = await question(rl, "Capability preset", defaults.preset);
         const agent = await question(rl, "Agent target", defaults.agent);
         const installSkills = locks.installSkills ?? (await yesNo(rl, "Install project-local skills now", defaults.installSkills));
-        const installMcpTools = locks.installMcpTools ?? (await yesNo(rl, "Run external MCP installers now", defaults.installMcpTools));
+        const installMcpTools = locks.installMcpTools ??
+            (await yesNo(rl, "Run finite external MCP tool installers now", defaults.installMcpTools));
         return { title, slug, packageName, preset, agent, installSkills, installMcpTools };
     }
     finally {

package/dist/src/stack.d.ts

@@ -8,13 +8,21 @@ export interface CapabilityPreset {
     mcp_servers: string[];
 }
 export interface McpServer {
+    readiness: string;
     priority: string;
+    execution_mode: string;
     source_need: string;
+    source: string;
+    hosted_url: string;
     install_command: string;
     uninstall_command: string;
+    setup_commands: string[];
     command: string;
     args: string[];
     env: Record<string, string>;
+    required_env: string[];
+    recommended_env: string[];
+    local_service: string;
     smoke_test: string;
     risks: string;
 }

package/dist/src/stack.js

@@ -20,38 +20,6 @@ export const AGENT_STACK = {
             commands: [
                 "npm exec --yes --package skills -- skills add existential-birds/beagle {agent_flag} --skill docling --copy -y"
             ]
-        },
-        optional_connectors: {
-            description: "CLI-side literature connectors to use only when MCP servers are unavailable.",
-            commands: [
-                "npm exec --yes --package skills -- skills add davila7/claude-code-templates {agent_flag} --skill openalex-database --copy -y",
-                "npm exec --yes --package skills -- skills add agents365-ai/365-skills {agent_flag} --skill semanticscholar-skill --copy -y",
-                "npm exec --yes --package skills -- skills add fuzhiyu/researchprojecttemplate {agent_flag} --skill zotero-paper-reader --copy -y"
-            ]
-        },
-        optional_mechanical_specialists: {
-            description: "Narrow mechanical helpers that do not replace project-native governance.",
-            commands: [
-                "npm exec --yes --package skills -- skills add bahayonghang/academic-writing-skills {agent_flag} --skill latex-paper-en --copy -y",
-                [
-                    "npm exec --yes --package skills -- skills add lllllllama/ai-paper-reproduction-skill",
-                    "{agent_flag}",
-                    "--skill",
-                    [
-                        "ai-research-reproduction",
-                        "repo-intake-and-plan",
-                        "env-and-assets-bootstrap",
-                        "minimal-run-and-audit",
-                        "paper-context-resolver",
-                        "analyze-project",
-                        "safe-debug",
-                        "run-train",
-                        "explore-run",
-                        "explore-code"
-                    ].join(" "),
-                    "--copy -y"
-                ].join(" ")
-            ]
         }
     },
     presets: {
@@ -61,119 +29,205 @@ export const AGENT_STACK = {
             mcp_servers: []
         },
         default: {
-            description: "Clean academic research setup with core scholarly MCP records.",
+            description: "Clean academic research setup with the low-friction arXiv MCP record.",
             skill_bundles: ["academic_research"],
-            mcp_servers: ["arxiv", "semantic-scholar", "openalex"]
+            mcp_servers: ["arxiv"]
         },
         enhanced: {
             description: "Default academic setup plus complementary agent engineering, document, frontend, testing, and doc conversion skills.",
             skill_bundles: ["academic_research", "default_complementary", "docling"],
-            mcp_servers: ["arxiv", "semantic-scholar", "openalex"]
+            mcp_servers: ["arxiv"]
         },
         literature: {
-            description: "Literature-heavy SOTA, survey, scoping, or systematic review setup.",
+            description: "Literature-heavy SOTA, survey, scoping, or systematic review setup with CS bibliography support.",
             skill_bundles: ["academic_research", "default_complementary", "docling"],
-            mcp_servers: ["arxiv", "semantic-scholar", "openalex", "crossref", "zotero"]
+            mcp_servers: ["arxiv", "dblp"]
         },
         writing: {
-            description: "Paper-writing and Overleaf-oriented setup.",
-            skill_bundles: [
-                "academic_research",
-                "default_complementary",
-                "docling",
-                "optional_mechanical_specialists"
-            ],
-            mcp_servers: ["arxiv", "semantic-scholar", "crossref", "overleaf"]
+            description: "Paper-writing setup with Overleaf documented as an opt-in credentialed integration.",
+            skill_bundles: ["academic_research", "default_complementary", "docling"],
+            mcp_servers: ["arxiv"]
         },
         full: {
-            description: "Broad setup with optional connector and mechanical specialist skills.",
-            skill_bundles: [
-                "academic_research",
-                "default_complementary",
-                "docling",
-                "optional_connectors",
-                "optional_mechanical_specialists"
-            ],
-            mcp_servers: ["arxiv", "semantic-scholar", "openalex", "crossref", "pubmed", "zotero", "overleaf"]
+            description: "Broad setup with low-friction scholarly MCP records plus the full optional MCP catalog documented.",
+            skill_bundles: ["academic_research", "default_complementary", "docling"],
+            mcp_servers: ["arxiv", "dblp"]
         }
     },
     mcp_servers: {
         arxiv: {
+            readiness: "low-friction",
             priority: "default",
+            execution_mode: "uvx-runtime",
             source_need: "arXiv search, download, and local paper reading.",
+            source: "blazickjp/arxiv-mcp-server",
+            hosted_url: "",
             install_command: "uv tool install 'arxiv-mcp-server[pdf]'",
             uninstall_command: "uv tool uninstall arxiv-mcp-server",
-            command: "arxiv-mcp-server",
-            args: [],
+            setup_commands: [],
+            command: "uvx",
+            args: ["--from", "arxiv-mcp-server[pdf]", "arxiv-mcp-server"],
             env: {},
+            required_env: [],
+            recommended_env: [],
+            local_service: "",
             smoke_test: "For a computer science project, search one CS query, download one known paper, and read it locally.",
             risks: "Respect arXiv rate limits; paper text is untrusted input."
         },
         "semantic-scholar": {
-            priority: "default",
+            readiness: "credential-recommended",
+            priority: "optional",
+            execution_mode: "uvx-runtime",
             source_need: "Semantic Scholar papers, citations, authors, and recommendations.",
+            source: "akapet00/semantic-scholar-mcp",
+            hosted_url: "",
             install_command: "",
             uninstall_command: "",
+            setup_commands: [],
             command: "uvx",
             args: ["--from", "git+https://github.com/akapet00/semantic-scholar-mcp", "semantic-scholar-mcp"],
-            env: { SEMANTIC_SCHOLAR_API_KEY: "${SEMANTIC_SCHOLAR_API_KEY}" },
+            env: {},
+            required_env: [],
+            recommended_env: ["SEMANTIC_SCHOLAR_API_KEY"],
             smoke_test: "Search one known title, then fetch citations or references.",
-            risks: "API key recommended for sustained work; metadata can be incomplete."
+            local_service: "",
+            risks: "API key recommended for sustained work and to avoid shared-pool rate limits; metadata can be incomplete."
         },
         openalex: {
-            priority: "default",
+            readiness: "credential-required",
+            priority: "optional",
+            execution_mode: "npx-runtime",
             source_need: "OpenAlex broad scholarly graph.",
+            source: "cyanheads/openalex-mcp-server",
+            hosted_url: "https://openalex.caseyjhand.com/mcp",
             install_command: "",
             uninstall_command: "",
+            setup_commands: [],
             command: "npx",
-            args: ["-y", "@cyanheads/openalex-mcp-server"],
-            env: { OPENALEX_API_KEY: "${OPENALEX_API_KEY_OR_EMAIL}" },
+            args: ["-y", "@cyanheads/openalex-mcp-server@latest"],
+            env: {},
+            required_env: ["OPENALEX_API_KEY"],
+            recommended_env: [],
+            local_service: "",
             smoke_test: "Search works by title or DOI and confirm stable OpenAlex IDs.",
-            risks: "Smoke-test before relying on it for final coverage."
+            risks: "The selected local server requires OPENALEX_API_KEY. OpenAlex keys are free for normal academic use with daily free usage; smoke-test coverage and cost headers before high-volume work."
         },
         crossref: {
+            readiness: "manual",
             priority: "manual",
+            execution_mode: "manual",
             source_need: "DOI and publication metadata.",
+            source: "manual selection required; candidate: AiAgentKarl/crossref-academic-mcp-server",
+            hosted_url: "",
             install_command: "",
             uninstall_command: "",
+            setup_commands: [],
             command: "",
             args: [],
             env: {},
+            required_env: [],
+            recommended_env: [],
+            local_service: "",
             smoke_test: "Resolve one DOI into publication metadata after choosing a maintained local server.",
-            risks: "Manual integration only; do not generate placeholder paths."
+            risks: "Manual integration only; current zero-friction Crossref-only MCP candidates are less mature than arXiv, DBLP, PubMed, or OpenAlex."
         },
         pubmed: {
+            readiness: "domain-specific",
             priority: "domain-specific",
+            execution_mode: "npx-runtime",
             source_need: "PubMed and biomedical literature.",
+            source: "cyanheads/pubmed-mcp-server",
+            hosted_url: "https://pubmed.caseyjhand.com/mcp",
             install_command: "",
             uninstall_command: "",
+            setup_commands: [],
             command: "npx",
-            args: ["-y", "@cyanheads/pubmed-mcp-server"],
-            env: { NCBI_API_KEY: "${NCBI_API_KEY}" },
+            args: ["-y", "@cyanheads/pubmed-mcp-server@latest"],
+            env: { MCP_TRANSPORT_TYPE: "stdio", MCP_LOG_LEVEL: "warning" },
+            required_env: [],
+            recommended_env: ["NCBI_API_KEY", "NCBI_ADMIN_EMAIL"],
+            local_service: "",
             smoke_test: "Search one PMID or title and fetch metadata.",
-            risks: "Domain-specific; observe NCBI rate limits."
+            risks: "Domain-specific; observe NCBI rate limits. API key and contact email improve reliability."
+        },
+        dblp: {
+            readiness: "low-friction-cs",
+            priority: "cs",
+            execution_mode: "uvx-runtime",
+            source_need: "DBLP computer science bibliography, venues, authors, and BibTeX.",
+            source: "szeider/mcp-dblp",
+            hosted_url: "",
+            install_command: "",
+            uninstall_command: "",
+            setup_commands: [],
+            command: "uvx",
+            args: ["mcp-dblp"],
+            env: {},
+            required_env: [],
+            recommended_env: [],
+            local_service: "",
+            smoke_test: "Search one known CS paper title and export or inspect its DBLP BibTeX.",
+            risks: "CS-specific metadata source; use with arXiv/Semantic Scholar/OpenAlex for coverage beyond DBLP."
         },
         zotero: {
+            readiness: "local-service",
             priority: "local-library",
+            execution_mode: "local-service",
             source_need: "Zotero local library and attachments.",
+            source: "eric-tramel/zoty",
+            hosted_url: "",
             install_command: "",
             uninstall_command: "",
+            setup_commands: ["uvx --refresh zoty setup", "uvx --refresh zoty doctor"],
             command: "uvx",
             args: ["zoty", "mcp"],
             env: {},
+            required_env: [],
+            recommended_env: [],
+            local_service: "Zotero desktop running with local API enabled; Zoty Bridge required for attachment and collection write operations.",
             smoke_test: "List collections, search one known item, and export BibTeX.",
             risks: "Requires Zotero local API and bridge setup."
         },
         overleaf: {
+            readiness: "manual-credentialed",
             priority: "writing",
+            execution_mode: "manual-local",
             source_need: "Overleaf project sync.",
-            install_command: "uv tool install overleaf-mcp-server",
-            uninstall_command: "uv tool uninstall overleaf-mcp-server",
-            command: "overleaf-mcp",
-            args: ["serve"],
-            env: { OVERLEAF_TOKEN: "${OVERLEAF_TOKEN}" },
+            source: "YounesBensafia/overleaf-mcp-server",
+            hosted_url: "",
+            install_command: "",
+            uninstall_command: "",
+            setup_commands: [],
+            command: "",
+            args: [],
+            env: {},
+            required_env: ["OVERLEAF_TOKEN"],
+            recommended_env: ["PROJECT_ID"],
+            local_service: "Local clone of the Overleaf MCP server configured with uv and an Overleaf project that supports Git sync.",
             smoke_test: "List projects or read one .tex file; do not write by default.",
-            risks: "Requires token setup; write access needs explicit approval."
+            risks: "Requires token and project setup; write/push access needs explicit approval."
+        },
+        "paper-search": {
+            readiness: "fallback",
+            priority: "fallback",
+            execution_mode: "manual-fallback",
+            source_need: "Multi-source paper search and download fallback across many scholarly sources.",
+            source: "openags/paper-search-mcp",
+            hosted_url: "",
+            install_command: "",
+            uninstall_command: "",
+            setup_commands: [],
+            command: "",
+            args: [],
+            env: {},
+            required_env: [],
+            recommended_env: [
+                "PAPER_SEARCH_MCP_UNPAYWALL_EMAIL",
+                "PAPER_SEARCH_MCP_SEMANTIC_SCHOLAR_API_KEY"
+            ],
+            local_service: "Manual review required before enabling; configure only permitted sources.",
+            smoke_test: "Search one harmless query with a source allow-list and verify provenance for each result.",
+            risks: "Powerful aggregator with optional restricted-source workflows; keep Sci-Hub or questionable download features disabled unless explicitly accepted."
         }
     }
 };

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "create-academic-research",
-  "version": "0.1.5",
+  "version": "0.1.7",
   "description": "Create and manage agent-ready academic research repositories.",
   "type": "module",
   "license": "MIT",

package/template/AGENTS.md

@@ -6,7 +6,8 @@ scholarly record.
 ## Core Rules
 
 - Preserve source originals under `sources/`, `data/raw/`, and `data/external/`.
-- Keep reusable logic in `src/`; keep notebooks for exploration and narrative output.
+- Keep reusable logic in `src/`; keep `scripts/` as thin repeatable entrypoints.
+- Keep notebooks for exploration and narrative output.
 - Tie claims to sources, datasets, experiment records, or decision records.
 - Update durable records when project knowledge changes.
 - Keep large data, generated caches, credentials, and private review material out of git.
@@ -35,6 +36,8 @@ scholarly record.
 - `wiki/synthesis.md` changes only when the project-level interpretation changes.
 - `wiki/open_questions.md` tracks unresolved questions.
 - `wiki/contradictions.md` tracks conflicts across sources, data, or runs.
+- `wiki/templates/` contains reusable source, claim, experiment, decision,
+  reviewer-concern, and research-question page structures.
 
 ## Evidence
 
@@ -45,3 +48,4 @@ scholarly record.
 - Citation issues go in `sources/bib/citation-audit.csv`.
 - SOTA records go in `sota/`.
 - Curated experiment records go in `experiments/`.
+- Repeatable command entrypoints go in `scripts/`.

package/template/README.md

@@ -41,6 +41,7 @@ npx academic-research doctor
 - `docs/agent/`: active agent workflows, capability profile, and MCP setup.
 - `docs/methodology/`: research design, evaluation plan, and validity threats.
 - `experiments/`: curated experiment registry and run records.
+- `scripts/`: thin repeatable entrypoints that call reusable code in `src/`.
 - `notebooks/`: optional exploratory and narrative notebooks.
 - `outputs/`: final figures, tables, models, and paper-supporting derived assets.
 - `reports/`: proposal, paper, slides, reviews, and rebuttal material.
@@ -58,19 +59,33 @@ npx academic-research skills install --preset default
 npx academic-research skills install --preset enhanced
 npx academic-research skills list
 npx academic-research skills status
-npx academic-research mcp enable arxiv semantic-scholar openalex
+npx academic-research setup
 npx academic-research mcp list
+npx academic-research mcp env openalex semantic-scholar zotero
+npx academic-research mcp enable arxiv dblp
 npx academic-research mcp commands arxiv
 npx academic-research mcp install arxiv
+npx academic-research mcp smoke
+npx academic-research mcp doctor
 ```
 
 `skills list` reports installed project-local skills. `skills presets` reports
 available install presets. `mcp enable` changes project records. `mcp commands`
-prints finite external install commands without running them. `mcp install`
-runs only finite tool installation commands; runtime-only `uvx`/`npx` MCP
-servers may have no install step and are started later by the MCP client.
+prints finite external install commands without running them. `mcp env` prints
+env vars, hosted endpoints, local prerequisites, and setup commands before you
+enable optional servers. `mcp install` runs only finite tool installation
+commands; runtime-only `uvx`/`npx` MCP servers may have no install step and are
+started later by the MCP client.
+
+`setup` prints the current project capability state, installed skill counts,
+enabled MCP records, and the next onboarding commands without changing files.
+`mcp smoke` performs a non-launching MCP readiness check: it reports required
+env vars, local/manual setup, and whether client runtime commands such as `uvx`
+or `npx` are available.
 
 `default` installs the companion academic research skill package and keeps the
-MCP records focused on core scholarly discovery. `enhanced` adds complementary
-external skills for agent engineering, frontend work, testing, document
-formats, and PDF conversion.
+MCP records focused on low-friction arXiv discovery. `literature` and `full`
+add DBLP for computer science bibliography. Credentialed, local-service, or
+domain-specific MCP servers such as OpenAlex, Semantic Scholar, PubMed, Zotero,
+and Overleaf should be enabled only after reading `docs/agent/mcp-setup.md` and
+checking their prerequisites with `mcp env`.

package/template/package.json

@@ -4,14 +4,17 @@
   "type": "module",
   "scripts": {
     "doctor": "academic-research doctor",
+    "setup": "academic-research setup",
     "skills:install": "academic-research skills install --preset default",
     "skills:list": "academic-research skills list",
     "skills:presets": "academic-research skills presets",
     "mcp:list": "academic-research mcp list",
     "mcp:commands": "academic-research mcp commands",
+    "mcp:env": "academic-research mcp env",
+    "mcp:smoke": "academic-research mcp smoke",
     "mcp:doctor": "academic-research mcp doctor"
   },
   "devDependencies": {
-    "create-academic-research": "0.1.5"
+    "create-academic-research": "0.1.7"
   }
 }

package/template/scripts/README.md

@@ -0,0 +1,10 @@
+# Scripts
+
+Use this folder for thin, repeatable command entrypoints.
+
+Scripts should orchestrate project code, not contain core scientific logic. Put
+reusable methods, models, data transforms, and analysis functions in `src/`.
+
+Each script should make its inputs, outputs, config path, and expected side
+effects clear enough to reproduce later from `docs/reproducibility/` or an
+experiment record.

package/template/tests/test_project_structure.py

@@ -8,6 +8,7 @@ def test_core_research_structure_exists() -> None:
         "configs/default.yaml",
         "configs/capabilities.yaml",
         "docs/agent/capability-profile.md",
+        "scripts/README.md",
         "notebooks/README.md",
         "outputs/figures",
         "outputs/tables",
@@ -21,5 +22,11 @@ def test_core_research_structure_exists() -> None:
         "sota/literature-matrix.csv",
         "wiki/index.md",
         "wiki/log.md",
+        "wiki/templates/source-page.md",
+        "wiki/templates/claim-page.md",
+        "wiki/templates/experiment-page.md",
+        "wiki/templates/decision-record.md",
+        "wiki/templates/reviewer-concern.md",
+        "wiki/templates/research-question.md",
     ):
         assert (root / relative).exists()

package/template/wiki/index.md

@@ -7,3 +7,12 @@ This is the content index for the LLM-maintained project wiki.
 - [[synthesis]]
 - [[open_questions]]
 - [[contradictions]]
+
+## Templates
+
+- [[templates/source-page]]
+- [[templates/claim-page]]
+- [[templates/experiment-page]]
+- [[templates/decision-record]]
+- [[templates/reviewer-concern]]
+- [[templates/research-question]]

package/template/wiki/templates/claim-page.md

@@ -0,0 +1,24 @@
+# Claim: <claim_id>
+
+## Claim
+
+<fill: atomic claim>
+
+## Status
+
+- Verdict:
+- Confidence:
+- Last checked:
+
+## Evidence
+
+| source_or_run_id | path | support_span | verdict | notes |
+|---|---|---|---|---|
+
+## Safe Wording
+
+<fill: wording that current evidence supports>
+
+## Risk
+
+<fill: what would weaken, falsify, or narrow this claim>

package/template/wiki/templates/decision-record.md

@@ -0,0 +1,26 @@
+# Decision: <decision_id>
+
+## Context
+
+<fill: situation and constraints>
+
+## Decision
+
+<fill: chosen option>
+
+## Alternatives Considered
+
+| option | reason rejected |
+|---|---|
+
+## Evidence
+
+- Sources:
+- Experiments:
+- Reviews:
+
+## Consequences
+
+- Expected benefit:
+- Risk:
+- Review date:

package/template/wiki/templates/experiment-page.md

@@ -0,0 +1,30 @@
+# Experiment: <run_id>
+
+## Question
+
+<fill: research question or hypothesis>
+
+## Run
+
+- Date:
+- Git commit:
+- Command:
+- Config:
+- Seed:
+- Dataset:
+- Metric:
+- Output path:
+
+## Result
+
+<fill: measured outcome>
+
+## Interpretation
+
+<fill: claim status, limitations, and next action>
+
+## Links
+
+- Registry row: `experiments/registry.csv`
+- Related claims:
+- Related sources:

package/template/wiki/templates/research-question.md

@@ -0,0 +1,27 @@
+# Research Question: <rq_id>
+
+## Question
+
+<fill: answerable research question>
+
+## Scope
+
+- Population/domain:
+- Evidence type:
+- Exclusions:
+
+## Contribution Link
+
+- Candidate claim:
+- Closest prior work:
+- Delta:
+- Failure condition:
+
+## Evidence Plan
+
+| evidence_needed | path_or_source | status | risk |
+|---|---|---|---|
+
+## Current Status
+
+<fill: hypothesis, active, answered, rejected, or out of scope>

package/template/wiki/templates/reviewer-concern.md

@@ -0,0 +1,25 @@
+# Reviewer Concern: <concern_id>
+
+## Concern
+
+<fill: reviewer or internal-review concern>
+
+## Classification
+
+- Source:
+- Severity:
+- Paper location:
+- Claim or issue:
+- Action: concede | clarify | add-evidence | reframe | defend | defer
+
+## Evidence
+
+<fill: source, experiment, or manuscript evidence>
+
+## Response Plan
+
+<fill: exact response and manuscript change>
+
+## Status
+
+<fill: open, addressed, deferred, rejected with rationale>

package/template/wiki/templates/source-page.md

@@ -0,0 +1,31 @@
+# Source: <source_id>
+
+## Bibliographic Record
+
+- Title:
+- Authors:
+- Year:
+- Venue:
+- Identifiers:
+- Native path:
+- Derived path:
+- Metadata path:
+
+## Summary
+
+<fill: concise source summary>
+
+## Evidence Record
+
+- Supports:
+- Contradicts:
+- Method / data / metric:
+- Limitations:
+- Project relevance:
+- Claim strength:
+- Allowed wording:
+- Forbidden stronger wording:
+
+## Follow-Up
+
+- <fill: unresolved checks>