@tangle-network/starter-foundry 0.7.2 → 0.13.0

package/registry/families/agent-runtime-physics-research-ts/files/methodology/literature-search.md

@@ -0,0 +1,113 @@
+---
+capability: literature-search
+status: active
+source: hand-authored, aligned with arXiv submission norms and APS / PRL conventions
+retrieved: 2026-04-26
+---
+
+# Literature Search methodology (physics)
+
+A literature search produces a `:::survey` block grounded in real
+papers retrieved via the research-corpus tools (arXiv, Semantic Scholar,
+OpenAlex, Crossref, INSPIRE-HEP for high-energy physics). Surveys are
+the input to citation analysis, experimental design, and theoretical
+position-taking — never the final deliverable.
+
+## When to use
+
+Trigger this template when the user asks:
+
+- "What's the state of the art on <topic>?"
+- "What experimental constraints exist on <observable>?"
+- "Find recent work on <method/regime/system>."
+- "Has anyone reproduced <result>?"
+- "What does the latest measurement say about <parameter>?"
+
+## Method
+
+1. **Clarify the regime.** Physics splits sharply across regimes:
+   classical / quantum, non-relativistic / relativistic, equilibrium /
+   non-equilibrium, weak / strong coupling, vacuum / condensed-matter.
+   The right corpus depends on the regime — INSPIRE-HEP for HEP and
+   astro-particle, arXiv `cond-mat` for condensed matter, ADS for
+   astronomy, etc.
+2. **Decompose the query into theory + system + observable + method.**
+   "Decoherence in NV centers measured via Ramsey" already names all
+   four; vague queries usually omit one. Ask before searching.
+3. **Pick corpora by regime.**
+   - Theory & broad physics: arXiv (`hep-th`, `quant-ph`, `cond-mat`,
+     `gr-qc`, `astro-ph`, `nucl-th`)
+   - HEP experiment, citation tracking: INSPIRE-HEP via OpenAlex
+   - Astronomy: ADS (NASA ADS) via Semantic Scholar bridge
+   - Cross-discipline / impact factors: Semantic Scholar, OpenAlex
+   - Last-resort or proceedings: Crossref
+4. **Build the query** — Boolean (`AND`, `OR`, `NOT`), arXiv category
+   filter (`cat:hep-th`), date range. Prefer canonical author names
+   when known (`Witten, E.` not `E. Witten`).
+5. **Filter for evidence quality and primary vs secondary.**
+   - Peer-reviewed (PRL, PRD, PRX, Nature Physics, JHEP) > arXiv
+     preprint > proceedings
+   - Recent measurement / first-principles calc > review > textbook
+   - Mark each tier in the output
+6. **Capture per-source.** Title, first author, venue, year,
+   theory/experiment/review tag, one-line contribution, one-line
+   relevance to the user's question, arXiv ID + DOI.
+7. **Synthesize.** A 2–4 sentence summary naming where the field
+   converges, where it remains contested, and what the next decisive
+   measurement / calculation would be.
+
+## Output shape
+
+```
+:::survey
+template: literature-search
+query: "<original user query>"
+query-date: <YYYY-MM-DD>
+corpora: [arxiv, openalex, inspire-hep]
+regime: { kind: "experiment", energy-scale: "TeV", system: "p-p collisions" }
+
+[surname, year]: <one-line contribution> [arXiv:XXXX.YYYYY | DOI:10.xxxx]
+[surname, year]: ...
+
+Synthesis: ...
+
+Open questions: ...
+:::
+```
+
+## Citation discipline
+
+Every claim of fact ships with a citation. Refuse to fabricate arXiv
+IDs, DOIs, author names, or measurement values. If the search returns
+nothing, say so; do not pad the survey with hallucinated entries.
+
+When the user asks for "the latest measurement of <parameter>," return
+the most recent published value with full uncertainty (statistical +
+systematic) and the experiment that measured it. PDG (Particle Data
+Group) is the canonical compilation for particle properties — cite the
+edition.
+
+## Common search-quality failures
+
+1. **arXiv-only tunnel.** arXiv is preprint-heavy and skews toward
+   theory/HEP. For applied/condensed-matter measurement work, hit
+   OpenAlex and the journal directly.
+2. **Date-window omitted.** Physics has slow-moving foundations (no
+   need to limit on textbook QED) and fast-moving experiments (Higgs,
+   gravitational-wave constraints update yearly). Specify the window.
+3. **Treating preprints as published.** arXiv submissions are
+   pre-publication. Always tag `arXiv-preprint` vs `published` in the
+   survey; some preprints never pass peer review.
+4. **Author-name ambiguity.** Common surnames (`Smith`, `Wang`,
+   `Zhang`) require disambiguation by ORCID, institution, or initials.
+5. **Missing PDG.** For particle masses, lifetimes, mixing angles —
+   start at the PDG, not arXiv.
+
+## Scope limits
+
+This template is a literature survey, not a meta-analysis. For a
+quantitative meta-analysis (combining measurements with proper error
+propagation), additionally specify: combination method (weighted
+average, BLUE, profile likelihood), correlated-uncertainty model, and
+external-error scaling. PDG documents its method explicitly — copy that
+discipline for any new compilation.

package/registry/families/agent-runtime-physics-research-ts/files/templates/citation-analysis.md

@@ -0,0 +1,37 @@
+# Citation Analysis Protocol
+
+## Purpose
+Analyze the citation network of a given paper or set of papers to understand influence, impact, and research fronts.
+
+## Steps
+1. **Identify target paper(s)**: Get the DOI, arxiv ID, or Semantic Scholar ID from the user.
+2. **Retrieve citation data**: Use Semantic Scholar's citation graph API to get citing and cited papers.
+3. **Compute metrics**:
+   - Citation count
+   - Citation velocity (citations per year)
+   - h-index (if multiple papers)
+   - Top citing authors/institutions
+4. **Visualize network**: Describe the citation network (e.g., "Paper A is cited by 15 papers, mostly in quantum optics; Paper B is a foundational reference with 200+ citations across condensed matter and quantum information").
+5. **Present results**: Emit a `:::survey` block with a summary table and network description.
+
+## Output format
+```
+:::survey
+template: citation-analysis
+target: [paper title/ID]
+
+**Citation Metrics**:
+- Total citations: ...
+- Citations per year: ...
+- Top citing fields: ...
+
+**Network Description**:
+[2-3 sentence description of the citation landscape]
+
+**Key Citing Papers**:
+- [surname, year] Title. Journal.
+:::
+```
+
+## Refusal
+If the paper is not found in any corpus, ask the user to verify the identifier. Do not fabricate citation data.

package/registry/families/agent-runtime-physics-research-ts/files/templates/experimental-design.md

@@ -0,0 +1,45 @@
+# Experimental Design Protocol
+
+## Purpose
+Help the user design a physics experiment with clear hypotheses, controls, and statistical considerations.
+
+## Steps
+1. **Define the research question**: What is the specific question the experiment aims to answer?
+2. **Formulate hypotheses**: State the null hypothesis (H0) and alternative hypothesis (H1) in precise, testable terms.
+3. **Identify variables**: Independent variable (what you manipulate), dependent variable (what you measure), controlled variables (what you keep constant).
+4. **Design the procedure**: Step-by-step description of the experimental setup, measurement techniques, and data collection.
+5. **Plan statistical analysis**: Specify the statistical test (e.g., t-test, chi-squared, Bayesian inference), significance level (e.g., α=0.05), and required sample size/power.
+6. **Address potential confounds**: List possible systematic errors, biases, and how to mitigate them.
+7. **Present the design**: Emit a `:::artifact` block with the full experimental design.
+
+## Output format
+```
+:::artifact
+template: experimental-design
+
+**Research Question**: ...
+
+**Hypotheses**:
+- H0: ...
+- H1: ...
+
+**Variables**:
+- Independent: ...
+- Dependent: ...
+- Controlled: ...
+
+**Procedure**:
+1. ...
+2. ...
+
+**Statistical Plan**:
+- Test: ...
+- α: ...
+- Sample size: ...
+
+**Confounds**: ...
+:::
+```
+
+## Refusal
+If the experiment involves human subjects, classified research, or requires ethical approval, escalate via `:::escalation` block.

package/registry/families/agent-runtime-physics-research-ts/files/templates/index.json

@@ -0,0 +1,25 @@
+{
+  "entries": [
+    {
+      "id": "literature-search",
+      "capability": "literature-search",
+      "path": "./literature-search.md",
+      "tags": ["methodology", "literature", "search", "arxiv"],
+      "version": "0.1.0"
+    },
+    {
+      "id": "citation-analysis",
+      "capability": "citation-analysis",
+      "path": "./citation-analysis.md",
+      "tags": ["methodology", "citation", "analysis", "semantic-scholar"],
+      "version": "0.1.0"
+    },
+    {
+      "id": "experimental-design",
+      "capability": "experimental-design",
+      "path": "./experimental-design.md",
+      "tags": ["methodology", "experiment", "design", "hypothesis"],
+      "version": "0.1.0"
+    }
+  ]
+}

package/registry/families/agent-runtime-physics-research-ts/files/templates/literature-search.md

@@ -0,0 +1,32 @@
+# Literature Search Protocol
+
+## Purpose
+Conduct a systematic literature search using the research-corpus tools (arxiv, Semantic Scholar, OpenAlex, Crossref) to find relevant papers on a given topic.
+
+## Steps
+1. **Clarify the query**: Ask the user to specify the research question, key concepts, and any constraints (date range, author, journal).
+2. **Construct search terms**: Use Boolean operators (AND, OR, NOT) and field-specific terms (e.g., "quantum entanglement AND decoherence").
+3. **Execute search**: Query each corpus tool. For arxiv, use the export API; for Semantic Scholar, use the graph API; for OpenAlex, use the works endpoint; for Crossref, use the works endpoint.
+4. **Deduplicate and rank**: Remove duplicates across sources. Rank by relevance (title/abstract match) and citation count.
+5. **Present results**: Emit a `:::survey` block with a table of papers (title, authors, year, source, citation count, link). Inline cite by `[surname, year]`.
+6. **Summarize**: Provide a brief synthesis of the key findings and gaps.
+
+## Output format
+```
+:::survey
+template: literature-search
+query: [original query]
+
+| # | Title | Authors | Year | Source | Citations | Link |
+|---|-------|---------|------|--------|-----------|------|
+| 1 | ... | ... | ... | arxiv | ... | ... |
+
+**Synthesis**: [2-3 sentence summary of main themes and gaps]
+
+**Full citations**:
+- [surname, year] Author, A. et al. (year). Title. Journal/arXiv. DOI
+:::
+```
+
+## Refusal
+If the query is too broad or vague, ask the user to narrow it. If no results found, state that clearly and suggest alternative terms.

package/registry/families/agent-runtime-physics-research-ts/files/wrangler.toml

@@ -0,0 +1,22 @@
+name = "physics-research"
+main = "src/index.ts"
+compatibility_date = "2025-03-01"
+
+[vars]
+AGENT_NAME = "physics-research"
+
+[[routes]]
+pattern = "/api/chat"
+methods = ["POST"]
+
+[[routes]]
+pattern = "/api/health"
+methods = ["GET"]
+
+[[env]]
+name = "TANGLE_API_KEY"
+value = ""
+
+[[env]]
+name = "AGENT_NAME"
+value = "physics-research"

package/registry/families/agent-runtime-physics-research-ts/manifest.json

@@ -0,0 +1,162 @@
+{
+  "id": "agent-runtime-physics-research-ts",
+  "description": "Physics research assistant agent bundle \u2014 literature search, citation-grounded analysis, and experimental design support. Composes the agent-runtime substrate with the research-corpus layer (arxiv / Semantic Scholar / OpenAlex / Crossref) so every claim is citation-grounded. Advisory only \u2014 never replaces peer review, ethics board, or professional judgment.",
+  "tags": [
+    "agent-runtime",
+    "physics-research",
+    "literature-search",
+    "citation-grounded",
+    "experimental-design",
+    "cloudflare-worker"
+  ],
+  "taxonomy": {
+    "language": "markdown",
+    "runtime": "cloudflare-worker",
+    "surface": "agent-runtime"
+  },
+  "includes": [
+    "agent-base:secure",
+    "agent-base:tangle",
+    "agent-tools:research-corpus",
+    "agent-output:blocks",
+    "agent-base:memory",
+    "agent-base:scheduler",
+    "agent-base:mcp-registry",
+    "agent-channels:telegram",
+    "agent-channels:discord",
+    "agent-channels:slack",
+    "agent-channels:whatsapp",
+    "agent-channels:imessage",
+    "agent-channels:gmail",
+    "agent-channels:linear"
+  ],
+  "defaults": {
+    "projectType": "agent-runtime-physics-research",
+    "agentName": "physics-research",
+    "allowedDomains": [
+      "api.tangle.tools",
+      "export.arxiv.org",
+      "api.semanticscholar.org",
+      "api.openalex.org",
+      "api.crossref.org"
+    ],
+    "allowedEnv": [
+      "TANGLE_API_KEY"
+    ],
+    "publicVars": [
+      "AGENT_NAME"
+    ],
+    "routes": [
+      {
+        "path": "/api/chat",
+        "auth": "bearer",
+        "callsTools": [
+          "research:paper-search",
+          "research:citation-graph"
+        ]
+      },
+      {
+        "path": "/api/health",
+        "auth": "none",
+        "callsTools": []
+      }
+    ],
+    "declaredCapabilities": [
+      "literature-search",
+      "citation-analysis",
+      "experimental-design"
+    ],
+    "runtime": {
+      "sandbox": "tangle",
+      "noNewPrivileges": true,
+      "capDrop": "ALL"
+    },
+    "stakes": "low",
+    "advisoryOnly": true
+  },
+  "files": [
+    {
+      "source": "files/agent.json",
+      "target": "agent.json"
+    },
+    {
+      "source": "files/AGENTS.md",
+      "target": "AGENTS.md"
+    },
+    {
+      "source": "files/templates/index.json",
+      "target": "templates/index.json"
+    },
+    {
+      "source": "files/templates/literature-search.md",
+      "target": "templates/literature-search.md"
+    },
+    {
+      "source": "files/templates/citation-analysis.md",
+      "target": "templates/citation-analysis.md"
+    },
+    {
+      "source": "files/templates/experimental-design.md",
+      "target": "templates/experimental-design.md"
+    },
+    {
+      "source": "files/methodology/index.json",
+      "target": "methodology/index.json"
+    },
+    {
+      "source": "files/methodology/literature-search.md",
+      "target": "methodology/literature-search.md"
+    },
+    {
+      "source": "files/methodology/citation-analysis.md",
+      "target": "methodology/citation-analysis.md"
+    },
+    {
+      "source": "files/methodology/experimental-design.md",
+      "target": "methodology/experimental-design.md"
+    },
+    {
+      "source": "files/wrangler.toml",
+      "target": "wrangler.toml"
+    },
+    {
+      "source": "files/README.md",
+      "target": "README.md"
+    }
+  ],
+  "contextHints": {
+    "commands": [],
+    "entrypoints": [
+      "AGENTS.md"
+    ],
+    "extensionPoints": [
+      "templates/",
+      "src/lib/tools/"
+    ]
+  },
+  "tieredKeywords": {
+    "tier1": [
+      "physics-research-agent",
+      "physics research agent",
+      "physics-researcher",
+      "physics researcher",
+      "literature-search-agent",
+      "literature search agent"
+    ],
+    "tier2": [
+      "arxiv",
+      "semantic-scholar",
+      "openalex",
+      "crossref",
+      "citation-graph",
+      "citation analysis",
+      "experimental-design",
+      "experimental design"
+    ],
+    "tier3": [],
+    "archetypes": [
+      "agent-in-sandbox",
+      "research-agent"
+    ]
+  }
+}

package/registry/families/agent-runtime-product-manager-ts/files/AGENTS.md

@@ -0,0 +1,58 @@
+---
+name: product-manager
+role: Product management advisor — product strategy, prioritization, discovery cycles. Not a substitute for the operator's product team or board.
+domain: product-mgmt
+allowedDomains:
+  - api.tangle.tools
+allowedEnv:
+  - TANGLE_API_KEY
+advisoryOnly: true
+escalationRequired: true
+version: 0.1.0
+---
+
+## Role
+
+You are a product management advisor. You work alongside a founder, product lead, or operator on three things: sharpening **product strategy**, running **prioritization frameworks**, and designing **discovery cycles** that reduce risk before building.
+
+You are **not a substitute for the operator's product team or board**. You do not see the cap table, the cash runway, the team's actual capacity, or the founder's relationship with the market. State this once, early, when the user asks for high-stakes calls (pivots, kill decisions, pricing).
+
+## Authoritative skills
+
+When the user's request maps to one of these capabilities, load the corresponding template *before* responding. The templates are the methodology source of truth; trust them over training.
+
+- `product-strategy` → `templates/product-strategy-canvas.md`
+- `prioritization-framework` → `templates/prioritization-framework.md`
+- `discovery-cycle` → `templates/discovery-cycle.md`
+
+## Output blocks
+
+Wrap structured deliverables in parseable blocks the host UI renders distinctly:
+
+- `:::artifact` — any persistent strategic output: a product strategy canvas, a prioritization matrix, a discovery plan
+- `:::analysis` — short interpretive readouts (e.g. "what this strategy loses you") that aren't the artifact itself but inform the user's next move
+
+Prose for conversational turns. Blocks only when there is a deliverable.
+
+## Refusal & escalation
+
+This is a moderate-stakes role; refusals are rare. Decline cleanly when:
+
+1. The user asks for legal, tax, or securities advice — redirect to counsel; do not approximate.
+2. The user asks you to make a decision they are accountable for (pivot/no-pivot, hire/fire, raise/no-raise) — name what you can contribute (the product surface of the question) and where the call belongs (the operator, the board, the cofounders).
+3. The user asks for competitive intelligence you don't have — design the experiment that would surface it instead.
+
+## What you WILL do
+
+- Pressure-test product strategy against the **Jobs-to-be-Done** lens (functional, social, emotional jobs; progress forces; anxiety forces).
+- Write **falsifiable** prioritization frameworks — every feature names a hypothesis, a primary metric, a sample-size proxy, and a kill criterion. No "let's see how it goes."
+- Push for **one** North-Star metric per product area. Refuse to bless a roadmap with five co-equal priorities.
+- Make the user articulate the **riskiest assumption** behind every initiative before building.
+- Name the trade-off behind every recommendation. A strategy that wins one segment loses three others; say which three.
+
+## What you WON'T do
+
+- Bless vague roadmaps ("build more features") without a hypothesis, metric, and kill criterion attached.
+- Recommend prioritization frameworks you can't justify against the user's stated strategy and resources.
+- Pretend to know market size, competitive intelligence, or customer willingness-to-pay you weren't given. If a number matters, ask the user for it or design the discovery cycle that would surface it.
+- Cosplay as a fractional PM with decision authority. You advise. The operator decides.

package/registry/families/agent-runtime-product-manager-ts/files/README.md

@@ -0,0 +1,30 @@
+# agent-runtime-product-manager-ts
+
+Product management agent bundle — helps operators define product strategy, prioritize features, and run discovery cycles.
+
+## Usage
+
+Deploy as a Cloudflare Worker. The agent exposes two endpoints:
+- `/api/chat` — main chat interface (bearer auth)
+- `/api/health` — health check (no auth)
+
+## Capabilities
+
+- **product-strategy**: Define and pressure-test product strategy using Jobs-to-be-Done.
+- **prioritization-framework**: Rank features with falsifiable hypotheses and kill criteria.
+- **discovery-cycle**: Run structured user research to reduce risk before building.
+
+## Configuration
+
+Set the following environment variables:
+- `TANGLE_API_KEY` — API key for Tangle router
+
+## Files
+
+- `system-prompt.md` — Agent system prompt
+- `templates/` — Methodology templates
+- `wrangler.toml` — Worker configuration
+
+## License
+
+MIT

package/registry/families/agent-runtime-product-manager-ts/files/TOOLS.md

@@ -0,0 +1,23 @@
+# TOOLS — agent-runtime-product-manager-ts
+
+The bundle inherits all primitives from `agent-base:secure`: `secrets`, `workspace`, `webhook-in`, `webhook-out`, `schedule`, `identity`, `audit`. Plus generic agent tools: `Read`, `Write`, `Edit`, `Glob`, `Grep`, `Bash`, `WebFetch`.
+
+This file lists **domain-specific tools the operator MAY add** if the deployment needs them. Each entry is intent — not implementation. The agent itself can build any of these on demand using `Bash`/`Write` if the operator hasn't.
+
+## Domain tools the operator may want
+
+Stub list — replace with role-specific entries. Keep ≤10 domain tools per deployment; more is fragmentation. Each tool should be:
+- Single-purpose
+- JSON output
+- ≤100 LOC
+- Listed here when added
+
+| Tool | Intent | Notes |
+|---|---|---|
+| `product-strategy` | Concrete tool implementing the `product-strategy` capability | Operator implements when needed |
+| `prioritization-framework` | Concrete tool implementing the `prioritization-framework` capability | Operator implements when needed |
+| `discovery-cycle` | Concrete tool implementing the `discovery-cycle` capability | Operator implements when needed |
+
+## Build a new tool when you need it
+
+The agent's job includes building tools when the measurement doesn't exist. New tools should be small, JSON-output, single-purpose. Add an entry here via PR after shipping.

package/registry/families/agent-runtime-product-manager-ts/files/agent.json

@@ -0,0 +1,39 @@
+{
+  "name": "product-manager",
+  "description": "Product management agent bundle — helps operators define product strategy, prioritize features, and run discovery cycles. Composes the agent-runtime substrate with structured-output-blocks on the Tangle floor. Advisory only — never replaces the operator's product team or board.",
+  "version": "0.1.0",
+  "tags": [
+    "agent-runtime",
+    "product-management",
+    "advisory-agent",
+    "product-strategy",
+    "prioritization",
+    "cloudflare-worker"
+  ],
+  "prompt": {
+    "systemPromptFile": "AGENTS.md"
+  },
+  "tools": {
+    "filesystem": true,
+    "shell": true,
+    "router": true
+  },
+  "permissions": {
+    "filesystem.write": "ask",
+    "shell.exec": "ask",
+    "network": "ask"
+  },
+  "resources": {
+    "files": [
+      {
+        "source": "methodology/"
+      },
+      {
+        "source": "README.md"
+      },
+      {
+        "source": "TOOLS.md"
+      }
+    ]
+  }
+}

package/registry/families/agent-runtime-product-manager-ts/files/methodology/discovery-cycle.md

@@ -0,0 +1,38 @@
+# Discovery Cycle
+
+## Overview
+A structured process to reduce risk before building by testing assumptions with real users.
+
+## Phases
+
+### 1. Frame
+- What is the decision we need to make?
+- What is the riskiest assumption?
+- What would we do differently if we knew the answer?
+
+### 2. Design
+- What method will we use? (interview, survey, prototype test, A/B test, data analysis)
+- Who will we talk to? (target customer segment, sample size)
+- What questions will we ask? (open-ended, avoid leading)
+
+### 3. Execute
+- Conduct the research. Record observations, not interpretations.
+- Capture verbatim quotes, behaviors, and emotional reactions.
+
+### 4. Analyze
+- What patterns emerged?
+- What did we learn about the assumption?
+- What new questions arose?
+
+### 5. Decide
+- Based on the evidence, do we proceed, pivot, or kill?
+- What is the next riskiest assumption to test?
+
+## Output
+A discovery brief with:
+- Decision made
+- Evidence summary
+- Next steps
+
+## Usage
+Run the cycle iteratively. Each cycle should take no more than 2 weeks. Start with the riskiest assumption.

package/registry/families/agent-runtime-product-manager-ts/files/methodology/index.json

@@ -0,0 +1,25 @@
+{
+  "entries": [
+    {
+      "id": "product-strategy-canvas",
+      "capability": "product-strategy",
+      "path": "./product-strategy-canvas.md",
+      "tags": ["methodology", "product-strategy", "jtbd"],
+      "version": "0.1.0"
+    },
+    {
+      "id": "prioritization-framework",
+      "capability": "prioritization-framework",
+      "path": "./prioritization-framework.md",
+      "tags": ["methodology", "prioritization", "roadmap"],
+      "version": "0.1.0"
+    },
+    {
+      "id": "discovery-cycle",
+      "capability": "discovery-cycle",
+      "path": "./discovery-cycle.md",
+      "tags": ["methodology", "discovery", "user-research"],
+      "version": "0.1.0"
+    }
+  ]
+}