The quickest path to identifying primary sources on a topic is to read widely in the secondary literature; any scholar writing about a development project or development issue will almost always make reference to specific reports in his/her notes and bibliography.  Take a look at this article as an example:  Rydzewski, Janusz R.  "Irrigation: A Viable Development Strategy?"  The Geographical Journal, Vol. 156, No. 2 (July 1990), pp. 175-180.

But, if you are already familiar with specific agencies or projects, then searching the web sites of intergovernmental organizations or non-profit organizations (e.g, UNDP, Engineers Without Borders) is another possible path.

Primary sources:

• Are original records created at the time an historical event occurred; they may also be written well after-the-fact by participants in the event. (Think of memoirs, autobiographies, or transcriptions of recollections - i.e. oral histories.)
• Need not focus on a specific event; rather, they may more generally reflect the ideas and values of a particular time period.
• May include personal writings such as letters, diaries, journals, memoirs, and autobiographies (in manuscript or published form). They may also include newspaper articles or editorials, speeches, interviews.
• May be produced by governments and organizations, for example, the World Bank, Doctors without Borders, Oxfam, etc.
• Need not be "non-fiction" works - or written works. They may be creative works, such as poems, novels, and plays, as well as photographs, audio recordings, moving pictures, sheet music, etc.

In short, primary sources serve as the raw material to interpret the past, and when they are used along with previous interpretations by historians, they provide the resources necessary for historical research.

Adapted from Using Primary Sources on the Web, History Section, Reference and User Services Association, American Library Association, accessed 9/22/11]